# Running on cnode7-008 # Started at Thu Nov 28 15:16:31 CST 2024 # SLURMD_NODENAME=cnode7-008 # SLURM_CLUSTER_NAME=slurm # SLURM_CONF=/cm/shared/apps/slurm/var/etc/slurm/slurm.conf # SLURM_CPUS_ON_NODE=224 # SLURM_CPUS_PER_TASK=8 # SLURM_EXPORT_ENV=PATH # SLURM_GET_USER_ENV=1 # SLURM_GPUS_ON_NODE=8 # SLURM_GPUS_PER_NODE=8 # SLURM_GTIDS=0 # SLURM_JOBID=4947 # SLURM_JOB_CPUS_PER_NODE='224(x2)' # SLURM_JOB_END_TIME=1764314188 # SLURM_JOB_GID=1026 # SLURM_JOB_GPUS=0,1,2,3,4,5,6,7 # SLURM_JOB_ID=4947 # SLURM_JOB_NAME=exp_owsm/s2t_train_05b_ds_raw_bpe50000/train.log # SLURM_JOB_NODELIST='cnode7-[008-009]' # SLURM_JOB_NUM_NODES=2 # SLURM_JOB_PARTITION=p2 # SLURM_JOB_QOS=normal # SLURM_JOB_START_TIME=1732778188 # SLURM_JOB_UID=1026 # SLURM_JOB_USER=williamchen # SLURM_LOCALID=0 # SLURM_MEM_PER_NODE=2048000 # SLURM_NNODES=2 # SLURM_NODEID=0 # SLURM_NODELIST='cnode7-[008-009]' # SLURM_NODE_ALIASES='(null)' # SLURM_OPEN_MODE=a # SLURM_PRIO_PROCESS=0 # SLURM_PROCID=0 # SLURM_SUBMIT_DIR=/mnt/home/williamchen/espnet/egs2/owsm_v3.1/s2t1 # SLURM_SUBMIT_HOST=154-T2-P1-NVR # SLURM_TASKS_PER_NODE='28(x2)' # SLURM_TASK_PID=1506698 # SLURM_TOPOLOGY_ADDR=cnode7-008 # SLURM_TOPOLOGY_ADDR_PATTERN=node # SLURM_WORKING_CLUSTER=slurm:154-T2-P1-NVR:6817:9984:109 # srun --export=ALL python3 -m espnet2.bin.s2t_train --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_v3/wav.scp,speech,kaldi_ark --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp_owsm/s2t_train_05b_ds_raw_bpe50000 --config conf/train_05b_ds.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp_owsm/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp,speech,kaldi_ark --train_shape_file exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev,text_prev,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc,text_ctc,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text,text,text --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev_v3/text.prev,text_prev,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_prev_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_ctc_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_shape --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///mnt/home/williamchen/espnet/egs2/owsm_v3.1/s2t1/exp_owsm/s2t_train_05b_ds_raw_bpe50000/.dist_init_ad8999db-5879-46f5-82be-ffc2b540bdc9 [2024-11-28 15:16:38,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:35,932] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/bin/python3 /mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_v3/wav.scp,speech,kaldi_ark --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp_owsm/s2t_train_05b_ds_raw_bpe50000 --config conf/train_05b_ds.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp_owsm/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp,speech,kaldi_ark --train_shape_file exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev,text_prev,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc,text_ctc,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text,text,text --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev_v3/text.prev,text_prev,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_prev_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_ctc_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_shape --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///mnt/home/williamchen/espnet/egs2/owsm_v3.1/s2t1/exp_owsm/s2t_train_05b_ds_raw_bpe50000/.dist_init_ad8999db-5879-46f5-82be-ffc2b540bdc9 /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/bin/python3 /mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_v3/wav.scp,speech,kaldi_ark --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/speech_shape --resume true --fold_length 80000 --output_dir exp_owsm/s2t_train_05b_ds_raw_bpe50000 --config conf/train_05b_ds.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp_owsm/s2t_stats_raw_bpe50000/train/feats_stats.npz --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp,speech,kaldi_ark --train_shape_file exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev,text_prev,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc,text_ctc,text --fold_length 150 --train_data_path_and_name_and_type exp_owsm/s2t_stats_raw_bpe50000/splits8/text,text,text --multiple_iterator true --valid_data_path_and_name_and_type dump/raw/dev_v3/text.prev,text_prev,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_prev_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_ctc_shape --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --valid_shape_file exp_owsm/s2t_stats_raw_bpe50000/valid/text_shape --ngpu 8 --multiprocessing_distributed true --dist_launcher slurm --dist_init_method file:///mnt/home/williamchen/espnet/egs2/owsm_v3.1/s2t1/exp_owsm/s2t_train_05b_ds_raw_bpe50000/.dist_init_ad8999db-5879-46f5-82be-ffc2b540bdc9 [2024-11-28 15:16:45,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:50,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:48,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:51,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:16:49,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [W1128 15:16:53.626546259 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:58.138527443 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:56.945336744 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.523730077 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:56.200340507 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.635364511 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.717878150 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:56.311098005 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:56.412043634 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.844165255 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.916484468 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.918638322 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:56.515050950 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [cnode7-008:0/16] 2024-11-28 15:16:59,598 (s2t:462) INFO: Vocabulary size: 50002 [W1128 15:16:56.613666123 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:57.678064114 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1128 15:16:59.194205814 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [cnode7-008:0/16] 2024-11-28 15:17:03,366 (abs_task:1383) INFO: pytorch.version=2.4.0+cu121, cuda.available=True, cudnn.version=90100, cudnn.benchmark=False, cudnn.deterministic=True [cnode7-008:0/16] 2024-11-28 15:17:03,371 (abs_task:1384) INFO: Model structure: ESPnetS2TModel( (frontend): DefaultFrontend( (stft): Stft(n_fft=512, win_length=400, hop_length=160, center=True, normalized=False, onesided=True) (frontend): Frontend() (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False) ) (specaug): SpecAug( (freq_mask): MaskAlongAxis(mask_width_range=[0, 27], num_mask=2, axis=freq) (time_mask): MaskAlongAxisVariableMaxWidth(mask_width_ratio_range=[0.0, 0.05], num_mask=10, axis=time) ) (normalize): GlobalMVN(stats_file=exp_owsm/s2t_stats_raw_bpe50000/train/feats_stats.npz, norm_means=True, norm_vars=True) (encoder): TransformerEncoder( (embed): Conv2dSubsampling( (conv): Sequential( (0): Conv2d(1, 1024, kernel_size=(3, 3), stride=(2, 2)) (1): ReLU() (2): Conv2d(1024, 1024, kernel_size=(3, 3), stride=(2, 2)) (3): ReLU() ) (out): Sequential( (0): Linear(in_features=19456, out_features=1024, bias=True) (1): PositionalEncoding( (dropout): Dropout(p=0.1, inplace=False) ) ) ) (encoders): MultiSequential( (0): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (1): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (2): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (3): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (4): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (5): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (6): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (7): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (8): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (9): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (10): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (11): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (12): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (13): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (14): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (15): EncoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) (after_norm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) ) (decoder): TransformerDecoder( (embed): Sequential( (0): Embedding(50002, 1024) (1): PositionalEncoding( (dropout): Dropout(p=0.1, inplace=False) ) ) (after_norm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (output_layer): Linear(in_features=1024, out_features=50002, bias=True) (decoders): MultiSequential( (0): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (1): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (2): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (3): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (4): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (5): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (6): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (7): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (8): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (9): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (10): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (11): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (12): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (13): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (14): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) (15): DecoderLayer( (self_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (src_attn): MultiHeadedAttention( (linear_q): Linear(in_features=1024, out_features=1024, bias=True) (linear_k): Linear(in_features=1024, out_features=1024, bias=True) (linear_v): Linear(in_features=1024, out_features=1024, bias=True) (linear_out): Linear(in_features=1024, out_features=1024, bias=True) (dropout): Identity() (q_norm): Identity() (k_norm): Identity() ) (feed_forward): PositionwiseFeedForward( (w_1): Linear(in_features=1024, out_features=4096, bias=True) (w_2): Linear(in_features=4096, out_features=1024, bias=True) (dropout): Dropout(p=0.1, inplace=False) (activation): ReLU() ) (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True) (dropout): Dropout(p=0.1, inplace=False) ) ) ) (criterion_att): LabelSmoothingLoss( (criterion): KLDivLoss() ) (ctc): CTC( (ctc_lo): Linear(in_features=1024, out_features=50002, bias=True) (ctc_loss): CTCLoss() ) ) Model summary: Class Name: ESPnetS2TModel Total Number of model parameters: 653.37 M Number of trainable parameters: 653.37 M (100.0%) Size: 2.61 GB Type: torch.float32 [cnode7-008:0/16] 2024-11-28 15:17:03,371 (abs_task:1387) INFO: Optimizer: Adadelta ( Parameter Group 0 capturable: False differentiable: False eps: 1e-06 foreach: None lr: 1.0 maximize: False rho: 0.9 weight_decay: 0 ) [cnode7-008:0/16] 2024-11-28 15:17:03,371 (abs_task:1388) INFO: Scheduler: None [cnode7-008:0/16] 2024-11-28 15:17:03,374 (abs_task:1397) INFO: Saving the configuration in exp_owsm/s2t_train_05b_ds_raw_bpe50000/config.yaml /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:03,273] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:03,275] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [cnode7-008:0/16] 2024-11-28 15:17:07,869 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:07,879] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:07,881] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:05,741] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:05,742] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:08,504] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:05,776] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:05,777] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-11-28 15:17:08,505] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:08,594] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:08,595] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:08,601] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:08,601] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-11-28 15:17:08,602] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:08,603] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:05,888] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:05,889] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [cnode7-008:0/16] 2024-11-28 15:17:08,721 (abs_task:1807) INFO: [valid] dataset: ESPnetDataset( speech: {"path": "dump/raw/dev_v3/wav.scp", "type": "kaldi_ark"} text_prev: {"path": "dump/raw/dev_v3/text.prev", "type": "text"} text_ctc: {"path": "dump/raw/dev_v3/text.ctc", "type": "text"} text: {"path": "dump/raw/dev_v3/text", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 15:17:08,721 (abs_task:1808) INFO: [valid] Batch sampler: UnsortedBatchSampler(N-batch=74743, batch_size=16, key_file=exp_owsm/s2t_stats_raw_bpe50000/valid/speech_shape, [cnode7-008:0/16] 2024-11-28 15:17:08,727 (abs_task:1809) INFO: [valid] mini-batch sizes summary: N-batch=74743, mean=16.0, min=16, max=17 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:08,898] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:08,899] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [cnode7-008:0/16] 2024-11-28 15:17:08,992 (distributed_utils:129) WARNING: ================================================================= Found OMP_NUM_THREADS=1 in environment variables. With some advanced features, DeepSpeed may have heavy cpu workload so that OMP_NUM_THREADS=1 is not sufficient. Try to increase it in your path.sh ================================================================= [2024-11-28 15:17:08,993] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:08,993] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed info: version=0.15.3, git-hash=unknown, git-branch=unknown [2024-11-28 15:17:08,993] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-008:1507505:1507505 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507505:1507505 [0] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507505:1507505 [0] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507505:1507505 [0] NCCL INFO cudaDriverVersion 12020 NCCL version 2.20.5+cuda12.4 cnode7-008:1507507:1507507 [2] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507507:1507507 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507507:1507507 [2] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507507:1507507 [2] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507511:1507511 [6] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507511:1507511 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507511:1507511 [6] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507511:1507511 [6] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507508:1507508 [3] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507508:1507508 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:06,449] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:06,451] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-009:637491:637491 [0] NCCL INFO cudaDriverVersion 12020 cnode7-009:637491:637491 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637491:637491 [0] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637491:637491 [0] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507509:1507509 [4] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507509:1507509 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507509:1507509 [4] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507509:1507509 [4] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-009:637497:637497 [6] NCCL INFO cudaDriverVersion 12020 cnode7-009:637497:637497 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637497:637497 [6] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637497:637497 [6] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507510:1507510 [5] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507510:1507510 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507510:1507510 [5] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507510:1507510 [5] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507512:1507512 [7] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507512:1507512 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507512:1507512 [7] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507512:1507512 [7] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-009:637492:637492 [1] NCCL INFO cudaDriverVersion 12020 cnode7-009:637492:637492 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637492:637492 [1] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637492:637492 [1] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507508:1507508 [3] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507508:1507508 [3] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-009:637498:637498 [7] NCCL INFO cudaDriverVersion 12020 cnode7-009:637498:637498 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637498:637498 [7] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637498:637498 [7] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:06,540] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:06,541] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-11-28 15:17:06,542] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:06,543] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-008:1507505:1508145 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507509:1508148 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637496:637496 [5] NCCL INFO cudaDriverVersion 12020 cnode7-009:637496:637496 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637496:637496 [5] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637496:637496 [5] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507510:1508149 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507507:1508146 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507512:1508151 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507511:1508147 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507508:1508150 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507505:1508145 [0] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507505:1508145 [0] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507505:1508145 [0] NCCL INFO Using network IB cnode7-009:637491:638091 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507509:1508148 [4] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507509:1508148 [4] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507509:1508148 [4] NCCL INFO Using network IB cnode7-009:637496:638097 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637495:637495 [4] NCCL INFO cudaDriverVersion 12020 cnode7-009:637495:637495 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637495:637495 [4] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637495:637495 [4] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-009:637493:637493 [2] NCCL INFO cudaDriverVersion 12020 cnode7-009:637493:637493 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637493:637493 [2] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637493:637493 [2] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-009:637497:638090 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507510:1508149 [5] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507510:1508149 [5] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507510:1508149 [5] NCCL INFO Using network IB cnode7-008:1507507:1508146 [2] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507507:1508146 [2] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507507:1508146 [2] NCCL INFO Using network IB cnode7-008:1507512:1508151 [7] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507512:1508151 [7] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507512:1508151 [7] NCCL INFO Using network IB cnode7-008:1507511:1508147 [6] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507511:1508147 [6] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507511:1508147 [6] NCCL INFO Using network IB cnode7-009:637492:638092 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507508:1508150 [3] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507508:1508150 [3] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507508:1508150 [3] NCCL INFO Using network IB cnode7-009:637498:638093 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637491:638091 [0] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637491:638091 [0] NCCL INFO Using non-device net plugin version 0 cnode7-009:637491:638091 [0] NCCL INFO Using network IB cnode7-009:637496:638097 [5] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637496:638097 [5] NCCL INFO Using non-device net plugin version 0 cnode7-009:637496:638097 [5] NCCL INFO Using network IB cnode7-009:637497:638090 [6] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637497:638090 [6] NCCL INFO Using non-device net plugin version 0 cnode7-009:637497:638090 [6] NCCL INFO Using network IB cnode7-009:637492:638092 [1] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637492:638092 [1] NCCL INFO Using non-device net plugin version 0 cnode7-009:637492:638092 [1] NCCL INFO Using network IB cnode7-009:637498:638093 [7] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637498:638093 [7] NCCL INFO Using non-device net plugin version 0 cnode7-009:637498:638093 [7] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:06,890] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:06,891] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:238: InstrumentationWarning: @typechecked only supports instrumenting functions wrapped with @classmethod, @staticmethod or @property -- not typechecking espnet2.train.deepspeed_trainer.DeepSpeedTrainer.valid_one_epoch def valid_one_epoch( [2024-11-28 15:17:09,981] [INFO] [comm.py:652:init_distributed] cdb=None [2024-11-28 15:17:09,983] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-009:637493:638103 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637495:638102 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637495:638102 [4] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637495:638102 [4] NCCL INFO Using non-device net plugin version 0 cnode7-009:637495:638102 [4] NCCL INFO Using network IB cnode7-009:637493:638103 [2] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637493:638103 [2] NCCL INFO Using non-device net plugin version 0 cnode7-009:637493:638103 [2] NCCL INFO Using network IB cnode7-009:637494:637494 [3] NCCL INFO cudaDriverVersion 12020 cnode7-009:637494:637494 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637494:637494 [3] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.171<0> cnode7-009:637494:637494 [3] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507506:1507506 [1] NCCL INFO cudaDriverVersion 12020 cnode7-008:1507506:1507506 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-008:1507506:1507506 [1] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.170<0> cnode7-008:1507506:1507506 [1] NCCL INFO NET/Plugin : dlerror=libnccl-net.so: cannot open shared object file: No such file or directory No plugin found (libnccl-net.so), using internal implementation cnode7-008:1507506:1508269 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637494:638363 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-009:637494:638363 [3] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.171<0> cnode7-009:637494:638363 [3] NCCL INFO Using non-device net plugin version 0 cnode7-009:637494:638363 [3] NCCL INFO Using network IB cnode7-008:1507506:1508269 [1] NCCL INFO NET/IB : Using [0]mlx5_0:1/IB [1]mlx5_3:1/IB [2]mlx5_4:1/IB [3]mlx5_5:1/IB [4]mlx5_6:1/IB [5]mlx5_7:1/RoCE [6]mlx5_8:1/IB [7]mlx5_9:1/IB [8]mlx5_10:1/IB [RO]; OOB enp170s0np0:10.225.1.170<0> cnode7-008:1507506:1508269 [1] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507506:1508269 [1] NCCL INFO Using network IB cnode7-009:637498:638093 [7] NCCL INFO comm 0x555580867bd0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507505:1508145 [0] NCCL INFO comm 0x5555647c9d00 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637494:638363 [3] NCCL INFO comm 0x55557c420950 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637491:638091 [0] NCCL INFO comm 0x555563b65950 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507509:1508148 [4] NCCL INFO comm 0x555572218f70 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637496:638097 [5] NCCL INFO comm 0x55557bc24510 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637495:638102 [4] NCCL INFO comm 0x555580e23030 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637493:638103 [2] NCCL INFO comm 0x555581023dc0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637497:638090 [6] NCCL INFO comm 0x55556ce1c820 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507510:1508149 [5] NCCL INFO comm 0x555560098520 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507506:1508269 [1] NCCL INFO comm 0x555560498dd0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507507:1508146 [2] NCCL INFO comm 0x55557c021be0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507512:1508151 [7] NCCL INFO comm 0x555567a24610 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507511:1508147 [6] NCCL INFO comm 0x55557541d130 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x29509662bfc31fb4 - Init START cnode7-009:637492:638092 [1] NCCL INFO comm 0x55558101da70 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507508:1508150 [3] NCCL INFO comm 0x55557c821fb0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x29509662bfc31fb4 - Init START cnode7-008:1507505:1508145 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507505:1508145 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-009:637497:638090 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637497:638090 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-008:1507511:1508147 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-009:637493:638103 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637493:638103 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-009:637496:638097 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637496:638097 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-009:637492:638092 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637492:638092 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-009:637498:638093 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637498:638093 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-009:637494:638363 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637494:638363 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-009:637491:638091 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637491:638091 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-009:637495:638102 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637495:638102 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-009:637491:638091 [0] NCCL INFO comm 0x555563b65950 rank 8 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-009:637491:638091 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 cnode7-009:637491:638091 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507509:1508148 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-008:1507509:1508148 [4] NCCL INFO comm 0x555572218f70 rank 4 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507509:1508148 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 cnode7-008:1507509:1508148 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507506:1508269 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507506:1508269 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-008:1507506:1508269 [1] NCCL INFO comm 0x555560498dd0 rank 1 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507506:1508269 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 cnode7-008:1507506:1508269 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507507:1508146 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507507:1508146 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-008:1507507:1508146 [2] NCCL INFO comm 0x55557c021be0 rank 2 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507507:1508146 [2] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507507:1508146 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 cnode7-008:1507507:1508146 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507512:1508151 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-008:1507512:1508151 [7] NCCL INFO comm 0x555567a24610 rank 7 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507512:1508151 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 cnode7-008:1507512:1508151 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637496:638097 [5] NCCL INFO comm 0x55557bc24510 rank 13 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-009:637496:638097 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 cnode7-009:637496:638097 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637495:638102 [4] NCCL INFO comm 0x555580e23030 rank 12 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-009:637495:638102 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 cnode7-009:637495:638102 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637498:638093 [7] NCCL INFO comm 0x555580867bd0 rank 15 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-009:637498:638093 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 cnode7-009:637498:638093 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637494:638363 [3] NCCL INFO comm 0x55557c420950 rank 11 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-009:637494:638363 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 cnode7-009:637494:638363 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637493:638103 [2] NCCL INFO comm 0x555581023dc0 rank 10 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-009:637493:638103 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 cnode7-009:637493:638103 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637497:638090 [6] NCCL INFO comm 0x55556ce1c820 rank 14 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-009:637497:638090 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 cnode7-009:637497:638090 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507510:1508149 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-008:1507510:1508149 [5] NCCL INFO comm 0x555560098520 rank 5 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507510:1508149 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 cnode7-008:1507510:1508149 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507511:1508147 [6] NCCL INFO comm 0x55557541d130 rank 6 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507511:1508147 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 cnode7-008:1507511:1508147 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507508:1508150 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507508:1508150 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-008:1507508:1508150 [3] NCCL INFO comm 0x55557c821fb0 rank 3 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507508:1508150 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 cnode7-008:1507508:1508150 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637492:638092 [1] NCCL INFO comm 0x55558101da70 rank 9 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-009:637492:638092 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 cnode7-009:637492:638092 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507505:1508145 [0] NCCL INFO comm 0x5555647c9d00 rank 0 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507505:1508145 [0] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 00/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 01/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 03/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 04/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 05/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 06/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 07/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 08/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 09/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 11/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 12/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 13/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 14/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 15/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-008:1507505:1508145 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 cnode7-008:1507505:1508145 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 12/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 14/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 10/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 08/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 08/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Connected all rings cnode7-009:637494:638363 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Connected all rings cnode7-009:637491:638091 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Connected all rings cnode7-008:1507509:1508148 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Connected all rings cnode7-009:637496:638097 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Connected all rings cnode7-009:637495:638102 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Connected all rings cnode7-009:637493:638103 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Connected all rings cnode7-009:637497:638090 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Connected all rings cnode7-008:1507506:1508269 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Connected all rings cnode7-008:1507510:1508149 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Connected all rings cnode7-008:1507507:1508146 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Connected all rings cnode7-008:1507511:1508147 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1508147 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Connected all rings cnode7-008:1507512:1508151 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Connected all rings cnode7-009:637492:638092 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Connected all rings cnode7-008:1507508:1508150 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Connected all rings cnode7-009:637498:638093 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:638093 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Connected all rings cnode7-008:1507505:1508145 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1508145 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:638363 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:638091 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1508148 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:638097 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:638102 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:638103 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:638090 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1508269 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-009:637492:638092 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1508149 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1508146 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1508151 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507508:1508150 [3] NCCL INFO Connected all trees cnode7-009:637494:638363 [3] NCCL INFO Connected all trees cnode7-008:1507507:1508146 [2] NCCL INFO Connected all trees cnode7-009:637493:638103 [2] NCCL INFO Connected all trees cnode7-008:1507507:1508146 [2] NCCL INFO NVLS comm 0x55557c021be0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507508:1508150 [3] NCCL INFO NVLS comm 0x55557c821fb0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637495:638102 [4] NCCL INFO Connected all trees cnode7-008:1507509:1508148 [4] NCCL INFO Connected all trees cnode7-008:1507510:1508149 [5] NCCL INFO Connected all trees cnode7-008:1507505:1508145 [0] NCCL INFO Connected all trees cnode7-008:1507512:1508151 [7] NCCL INFO Connected all trees cnode7-008:1507506:1508269 [1] NCCL INFO Connected all trees cnode7-009:637496:638097 [5] NCCL INFO Connected all trees cnode7-009:637497:638090 [6] NCCL INFO Connected all trees cnode7-008:1507511:1508147 [6] NCCL INFO Connected all trees cnode7-009:637498:638093 [7] NCCL INFO Connected all trees cnode7-009:637492:638092 [1] NCCL INFO Connected all trees cnode7-009:637491:638091 [0] NCCL INFO Connected all trees cnode7-008:1507505:1508145 [0] NCCL INFO NVLS comm 0x5555647c9d00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637494:638363 [3] NCCL INFO NVLS comm 0x55557c420950 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637491:638091 [0] NCCL INFO NVLS comm 0x555563b65950 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507509:1508148 [4] NCCL INFO NVLS comm 0x555572218f70 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637496:638097 [5] NCCL INFO NVLS comm 0x55557bc24510 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637495:638102 [4] NCCL INFO NVLS comm 0x555580e23030 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637497:638090 [6] NCCL INFO NVLS comm 0x55556ce1c820 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637493:638103 [2] NCCL INFO NVLS comm 0x555581023dc0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507506:1508269 [1] NCCL INFO NVLS comm 0x555560498dd0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507510:1508149 [5] NCCL INFO NVLS comm 0x555560098520 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507511:1508147 [6] NCCL INFO NVLS comm 0x55557541d130 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507512:1508151 [7] NCCL INFO NVLS comm 0x555567a24610 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637492:638092 [1] NCCL INFO NVLS comm 0x55558101da70 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637498:638093 [7] NCCL INFO NVLS comm 0x555580867bd0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507505:1508145 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:638093 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1508145 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:638363 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:638091 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1508148 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:638097 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:638102 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:638090 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1508269 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1508149 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1508147 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1508146 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1508151 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:638092 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1508150 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-009:637493:638103 [2] NCCL INFO Connected NVLS tree cnode7-009:637493:638103 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637493:638103 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507506:1508269 [1] NCCL INFO Connected NVLS tree cnode7-008:1507506:1508269 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507506:1508269 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507507:1508146 [2] NCCL INFO Connected NVLS tree cnode7-008:1507507:1508146 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507507:1508146 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637492:638092 [1] NCCL INFO Connected NVLS tree cnode7-009:637492:638092 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637492:638092 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507508:1508150 [3] NCCL INFO Connected NVLS tree cnode7-008:1507508:1508150 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507508:1508150 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637498:638093 [7] NCCL INFO Connected NVLS tree cnode7-009:637498:638093 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637498:638093 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507505:1508145 [0] NCCL INFO Connected NVLS tree cnode7-008:1507505:1508145 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507505:1508145 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507510:1508149 [5] NCCL INFO Connected NVLS tree cnode7-008:1507510:1508149 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507510:1508149 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507512:1508151 [7] NCCL INFO Connected NVLS tree cnode7-008:1507512:1508151 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507512:1508151 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507511:1508147 [6] NCCL INFO Connected NVLS tree cnode7-008:1507511:1508147 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507511:1508147 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637494:638363 [3] NCCL INFO Connected NVLS tree cnode7-009:637494:638363 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637494:638363 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637491:638091 [0] NCCL INFO Connected NVLS tree cnode7-009:637491:638091 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637491:638091 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507509:1508148 [4] NCCL INFO Connected NVLS tree cnode7-008:1507509:1508148 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507509:1508148 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637496:638097 [5] NCCL INFO Connected NVLS tree cnode7-009:637496:638097 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637496:638097 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637495:638102 [4] NCCL INFO Connected NVLS tree cnode7-009:637495:638102 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637495:638102 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637497:638090 [6] NCCL INFO Connected NVLS tree cnode7-009:637497:638090 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637497:638090 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507511:1508147 [6] NCCL INFO comm 0x55557541d130 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507507:1508146 [2] NCCL INFO comm 0x55557c021be0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507512:1508151 [7] NCCL INFO comm 0x555567a24610 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507510:1508149 [5] NCCL INFO comm 0x555560098520 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507506:1508269 [1] NCCL INFO comm 0x555560498dd0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507505:1508145 [0] NCCL INFO comm 0x5555647c9d00 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507508:1508150 [3] NCCL INFO comm 0x55557c821fb0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-008:1507509:1508148 [4] NCCL INFO comm 0x555572218f70 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637498:638093 [7] NCCL INFO comm 0x555580867bd0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637494:638363 [3] NCCL INFO comm 0x55557c420950 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637497:638090 [6] NCCL INFO comm 0x55556ce1c820 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637492:638092 [1] NCCL INFO comm 0x55558101da70 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637493:638103 [2] NCCL INFO comm 0x555581023dc0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637495:638102 [4] NCCL INFO comm 0x555580e23030 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637491:638091 [0] NCCL INFO comm 0x555563b65950 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x29509662bfc31fb4 - Init COMPLETE cnode7-009:637496:638097 [5] NCCL INFO comm 0x55557bc24510 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x29509662bfc31fb4 - Init COMPLETE [2024-11-28 15:17:25,625] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False [2024-11-28 15:17:25,635] [INFO] [logging.py:129:log_dist] [Rank 0] Using DeepSpeed Optimizer param name adam as basic optimizer [2024-11-28 15:17:25,635] [INFO] [logging.py:129:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer [2024-11-28 15:17:25,669] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Basic Optimizer = FusedAdam [2024-11-28 15:17:25,669] [INFO] [utils.py:59:is_zero_supported_optimizer] Checking ZeRO support for optimizer=FusedAdam type= [2024-11-28 15:17:25,669] [INFO] [logging.py:129:log_dist] [Rank 0] Creating torch.bfloat16 ZeRO stage 2 optimizer [2024-11-28 15:17:25,669] [INFO] [stage_1_and_2.py:149:__init__] Reduce bucket size 500000000 [2024-11-28 15:17:25,669] [INFO] [stage_1_and_2.py:150:__init__] Allgather bucket size 500000000 [2024-11-28 15:17:25,669] [INFO] [stage_1_and_2.py:151:__init__] CPU Offload: False [2024-11-28 15:17:25,669] [INFO] [stage_1_and_2.py:152:__init__] Round robin gradient partitioning: False [2024-11-28 15:17:27,393] [INFO] [utils.py:781:see_memory_usage] Before initializing optimizer states [2024-11-28 15:17:27,396] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.45 GB CA 1.45 GB Max_CA 1 GB [2024-11-28 15:17:27,397] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 72.66 GB, percent = 3.6% [2024-11-28 15:17:27,515] [INFO] [utils.py:781:see_memory_usage] After initializing optimizer states [2024-11-28 15:17:27,516] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.52 GB CA 1.6 GB Max_CA 2 GB [2024-11-28 15:17:27,516] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 71.76 GB, percent = 3.6% [2024-11-28 15:17:27,516] [INFO] [stage_1_and_2.py:544:__init__] optimizer state initialized [2024-11-28 15:17:27,627] [INFO] [utils.py:781:see_memory_usage] After initializing ZeRO optimizer [2024-11-28 15:17:27,627] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.37 GB CA 1.6 GB Max_CA 2 GB [2024-11-28 15:17:27,627] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 70.62 GB, percent = 3.5% [2024-11-28 15:17:27,629] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Final Optimizer = DeepSpeedZeroOptimizer [2024-11-28 15:17:27,629] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed using client callable to create LR scheduler [2024-11-28 15:17:27,629] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed LR Scheduler = PiecewiseLinearWarmupLR(warmup_steps_list=[0, 30000, 60000], warmup_lr_list=[0.0, 5e-05, 0.0002]) [2024-11-28 15:17:27,629] [INFO] [logging.py:129:log_dist] [Rank 0] step=0, skipped=0, lr=[np.float64(1.6666666666666667e-09)], mom=[[0.9, 0.98]] [2024-11-28 15:17:27,630] [INFO] [config.py:999:print] DeepSpeedEngine configuration: [2024-11-28 15:17:27,630] [INFO] [config.py:1003:print] activation_checkpointing_config { "partition_activations": false, "contiguous_memory_optimization": false, "cpu_checkpointing": false, "number_checkpoints": null, "synchronize_checkpoint_boundary": false, "profile": false } [2024-11-28 15:17:27,630] [INFO] [config.py:1003:print] aio_config ................... {'block_size': 1048576, 'queue_depth': 8, 'thread_count': 1, 'single_submit': False, 'overlap_events': True, 'use_gds': False} [2024-11-28 15:17:27,630] [INFO] [config.py:1003:print] amp_enabled .................. False [2024-11-28 15:17:27,630] [INFO] [config.py:1003:print] amp_params ................... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] autotuning_config ............ { "enabled": false, "start_step": null, "end_step": null, "metric_path": null, "arg_mappings": null, "metric": "throughput", "model_info": null, "results_dir": "autotuning_results", "exps_dir": "autotuning_exps", "overwrite": true, "fast": true, "start_profile_step": 3, "end_profile_step": 5, "tuner_type": "gridsearch", "tuner_early_stopping": 5, "tuner_num_trials": 50, "model_info_path": null, "mp_size": 1, "max_train_batch_size": null, "min_train_batch_size": 1, "max_train_micro_batch_size_per_gpu": 1.024000e+03, "min_train_micro_batch_size_per_gpu": 1, "num_tuning_micro_batch_sizes": 3 } [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] bfloat16_enabled ............. True [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] bfloat16_immediate_grad_update False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] checkpoint_parallel_write_pipeline False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] checkpoint_tag_validation_enabled True [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] checkpoint_tag_validation_fail False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] comms_config ................. [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] communication_data_type ...... None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] compression_config ........... {'weight_quantization': {'shared_parameters': {'enabled': False, 'quantizer_kernel': False, 'schedule_offset': 0, 'quantize_groups': 1, 'quantize_verbose': False, 'quantization_type': 'symmetric', 'quantize_weight_in_forward': False, 'rounding': 'nearest', 'fp16_mixed_quantize': False, 'quantize_change_ratio': 0.001}, 'different_groups': {}}, 'activation_quantization': {'shared_parameters': {'enabled': False, 'quantization_type': 'symmetric', 'range_calibration': 'dynamic', 'schedule_offset': 1000}, 'different_groups': {}}, 'sparse_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'row_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'head_pruning': {'shared_parameters': {'enabled': False, 'method': 'topk', 'schedule_offset': 1000}, 'different_groups': {}}, 'channel_pruning': {'shared_parameters': {'enabled': False, 'method': 'l1', 'schedule_offset': 1000}, 'different_groups': {}}, 'layer_reduction': {'enabled': False}} [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] curriculum_enabled_legacy .... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] curriculum_params_legacy ..... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] data_efficiency_config ....... {'enabled': False, 'seed': 1234, 'data_sampling': {'enabled': False, 'num_epochs': 1000, 'num_workers': 0, 'curriculum_learning': {'enabled': False}}, 'data_routing': {'enabled': False, 'random_ltd': {'enabled': False, 'layer_token_lr_schedule': {'enabled': False}}}} [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] data_efficiency_enabled ...... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] dataloader_drop_last ......... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] disable_allgather ............ False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] dump_state ................... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] dynamic_loss_scale_args ...... None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_enabled ........... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_gas_boundary_resolution 1 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_layer_name ........ bert.encoder.layer [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_layer_num ......... 0 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_max_iter .......... 100 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_stability ......... 1e-06 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_tol ............... 0.01 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] eigenvalue_verbose ........... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] elasticity_enabled ........... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] flops_profiler_config ........ { "enabled": false, "recompute_fwd_factor": 0.0, "profile_step": 1, "module_depth": -1, "top_modules": 1, "detailed": true, "output_file": null } [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] fp16_auto_cast ............... None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] fp16_enabled ................. False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] fp16_master_weights_and_gradients False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] global_rank .................. 0 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] grad_accum_dtype ............. None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] gradient_accumulation_steps .. 1 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] gradient_clipping ............ 5.0 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] gradient_predivide_factor .... 1.0 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] graph_harvesting ............. False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] hybrid_engine ................ enabled=False max_out_tokens=512 inference_tp_size=1 release_inference_cache=False pin_parameters=True tp_gather_partition_size=8 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] initial_dynamic_scale ........ 1 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] load_universal_checkpoint .... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] loss_scale ................... 1.0 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] memory_breakdown ............. False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] mics_hierarchial_params_gather False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] mics_shard_size .............. -1 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] monitor_config ............... tensorboard=TensorBoardConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') comet=CometConfig(enabled=False, samples_log_interval=100, project=None, workspace=None, api_key=None, experiment_name=None, experiment_key=None, online=None, mode=None) wandb=WandbConfig(enabled=False, group=None, team=None, project='deepspeed') csv_monitor=CSVConfig(enabled=False, output_path='', job_name='DeepSpeedJobName') [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] nebula_config ................ { "enabled": false, "persistent_storage_path": null, "persistent_time_interval": 100, "num_of_version_in_retention": 2, "enable_nebula_load": true, "load_path": null } [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] optimizer_legacy_fusion ...... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] optimizer_name ............... adam [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] optimizer_params ............. {'lr': 0.0002, 'betas': [0.9, 0.98], 'eps': 1e-06, 'weight_decay': 0.0} [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True} [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] pld_enabled .................. False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] pld_params ................... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] prescale_gradients ........... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] scheduler_name ............... None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] scheduler_params ............. None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] seq_parallel_communication_data_type torch.float32 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] sparse_attention ............. None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] sparse_gradients_enabled ..... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] steps_per_print .............. 1000 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] timers_config ................ enabled=True synchronized=True [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] train_batch_size ............. 16 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] train_micro_batch_size_per_gpu 1 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] use_data_before_expert_parallel_ False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] use_node_local_storage ....... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] wall_clock_breakdown ......... False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] weight_quantization_config ... None [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] world_size ................... 16 [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] zero_allow_untested_optimizer False [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] zero_config .................. stage=2 contiguous_gradients=True reduce_scatter=True reduce_bucket_size=500000000 use_multi_rank_bucket_allreduce=True allgather_partitions=True allgather_bucket_size=500000000 overlap_comm=True load_from_fp32_weights=True elastic_checkpoint=False offload_param=None offload_optimizer=None sub_group_size=1000000000 cpu_offload_param=None cpu_offload_use_pin_memory=None cpu_offload=None prefetch_bucket_size=50000000 param_persistence_threshold=100000 model_persistence_threshold=9223372036854775807 max_live_parameters=1000000000 max_reuse_distance=1000000000 gather_16bit_weights_on_model_save=False use_all_reduce_for_fetch_params=False stage3_gather_fp16_weights_on_model_save=False ignore_unused_parameters=True legacy_stage1=False round_robin_gradients=False zero_hpz_partition_size=1 zero_quantized_weights=False zero_quantized_nontrainable_weights=False zero_quantized_gradients=False mics_shard_size=-1 mics_hierarchical_params_gather=False memory_efficient_linear=True pipeline_loading_checkpoint=False override_module_apply=True [2024-11-28 15:17:27,631] [INFO] [config.py:1003:print] zero_enabled ................. True [2024-11-28 15:17:27,632] [INFO] [config.py:1003:print] zero_force_ds_cpu_optimizer .. True [2024-11-28 15:17:27,632] [INFO] [config.py:1003:print] zero_optimization_stage ...... 2 [2024-11-28 15:17:27,632] [INFO] [config.py:989:print_user_config] json = { "train_micro_batch_size_per_gpu": 1, "gradient_accumulation_steps": 1, "gradient_clipping": 5.0, "bf16": { "enabled": true }, "zero_optimization": { "stage": 2, "contiguous_gradients": true, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 5.000000e+08, "allgather_bucket_size": 5.000000e+08 }, "optimizer": { "type": "Adam", "params": { "lr": 0.0002, "betas": [0.9, 0.98], "eps": 1e-06, "weight_decay": 0.0 } }, "wall_clock_breakdown": false, "steps_per_print": 1000 } [cnode7-008:0/16] 2024-11-28 15:17:27,633 (deepspeed_trainer:70) INFO: Try to resume but find no checkpoint [cnode7-008:0/16] 2024-11-28 15:17:27,635 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-28 15:17:52,967 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 15:17:58,220 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 15:17:58,221 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-28 15:17:58,223 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 15:18:20,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:20,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:18,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:19,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:23,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:21,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:21,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:25,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:25,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:25,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:25,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:22,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:25,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:23,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:23,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:18:23,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:06,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:05,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:05,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:09,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:10,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:10,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:10,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:16,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:13,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:16,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:22,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:51,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:51,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:52,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:57,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:00,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:19:57,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:00,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:03,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:04,312] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:01,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:02,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:02,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:05,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:03,154] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:05,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:10,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:38,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:38,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:40,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:46,591] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:44,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:45,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:50,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:53,627] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:53,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:51,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:54,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:56,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:53,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:57,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:20:56,532] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507505:1515283 [0] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507505:1515283 [0] NCCL INFO Using network IB [2024-11-28 15:21:03,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-009:637495:645505 [4] NCCL INFO Using non-device net plugin version 0 cnode7-009:637495:645505 [4] NCCL INFO Using network IB cnode7-009:637494:645506 [3] NCCL INFO Using non-device net plugin version 0 cnode7-009:637494:645506 [3] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507512:1515700 [7] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507512:1515700 [7] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-009:637496:645919 [5] NCCL INFO Using non-device net plugin version 0 cnode7-009:637496:645919 [5] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-009:637498:646011 [7] NCCL INFO Using non-device net plugin version 0 cnode7-009:637498:646011 [7] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507507:1516100 [2] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507507:1516100 [2] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507510:1516380 [5] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507510:1516380 [5] NCCL INFO Using network IB cnode7-008:1507506:1516381 [1] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507506:1516381 [1] NCCL INFO Using network IB cnode7-009:637497:646298 [6] NCCL INFO Using non-device net plugin version 0 cnode7-009:637497:646298 [6] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-009:637493:646472 [2] NCCL INFO Using non-device net plugin version 0 cnode7-009:637493:646472 [2] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507511:1516731 [6] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507511:1516731 [6] NCCL INFO Using network IB cnode7-008:1507508:1516799 [3] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507508:1516799 [3] NCCL INFO Using network IB cnode7-008:1507509:1516800 [4] NCCL INFO Using non-device net plugin version 0 cnode7-008:1507509:1516800 [4] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-009:637492:646688 [1] NCCL INFO Using non-device net plugin version 0 cnode7-009:637492:646688 [1] NCCL INFO Using network IB /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): cnode7-008:1507508:1516799 [3] NCCL INFO comm 0x5555c2bd3d50 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507505:1515283 [0] NCCL INFO comm 0x5555b8bbaf30 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637492:646688 [1] NCCL INFO comm 0x5555c2bf3600 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637498:646011 [7] NCCL INFO comm 0x5555d6bd2ca0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507506:1516381 [1] NCCL INFO comm 0x5555c2c00610 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507509:1516800 [4] NCCL INFO comm 0x5555c2bf6b30 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507511:1516731 [6] NCCL INFO comm 0x5555c2bc8c60 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507510:1516380 [5] NCCL INFO comm 0x5555c2bccdd0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507507:1516100 [2] NCCL INFO comm 0x5555c2bf72f0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637495:645505 [4] NCCL INFO comm 0x5555c2c27fe0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637491:647053 [0] NCCL INFO Using non-device net plugin version 0 cnode7-009:637491:647053 [0] NCCL INFO Using network IB cnode7-009:637491:647053 [0] NCCL INFO comm 0x5555b8bd9f40 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637493:646472 [2] NCCL INFO comm 0x5555c2c0bbd0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xebe7a4df1cc421c2 - Init START cnode7-008:1507512:1515700 [7] NCCL INFO comm 0x5555c2be6e00 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637497:646298 [6] NCCL INFO comm 0x5555d6c24420 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637494:645506 [3] NCCL INFO comm 0x5555c2bfa560 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637496:645919 [5] NCCL INFO comm 0x5555c2bf8710 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xebe7a4df1cc421c2 - Init START cnode7-009:637493:646472 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637493:646472 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-009:637497:646298 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637497:646298 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-009:637496:645919 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637496:645919 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-009:637498:646011 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637498:646011 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-009:637492:646688 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637492:646688 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-008:1507509:1516800 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-009:637491:647053 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637491:647053 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-009:637495:645505 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-009:637495:645505 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-009:637494:645506 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-009:637494:645506 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-008:1507508:1516799 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507508:1516799 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-008:1507505:1515283 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507505:1515283 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-008:1507506:1516381 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507506:1516381 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-008:1507506:1516381 [1] NCCL INFO comm 0x5555c2c00610 rank 1 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507506:1516381 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 cnode7-008:1507506:1516381 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507505:1515283 [0] NCCL INFO comm 0x5555b8bbaf30 rank 0 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 00/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 01/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 03/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 04/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 05/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 06/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 07/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 08/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 09/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 11/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 12/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 13/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 14/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-008:1507505:1515283 [0] NCCL INFO Channel 15/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-008:1507505:1515283 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 cnode7-008:1507505:1515283 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507512:1515700 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-008:1507512:1515700 [7] NCCL INFO comm 0x5555c2be6e00 rank 7 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507512:1515700 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 cnode7-008:1507512:1515700 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507507:1516100 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-008:1507507:1516100 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-008:1507507:1516100 [2] NCCL INFO comm 0x5555c2bf72f0 rank 2 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507507:1516100 [2] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507507:1516100 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 cnode7-008:1507507:1516100 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507508:1516799 [3] NCCL INFO comm 0x5555c2bd3d50 rank 3 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507508:1516799 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 cnode7-008:1507508:1516799 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637491:647053 [0] NCCL INFO comm 0x5555b8bd9f40 rank 8 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-009:637491:647053 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 cnode7-009:637491:647053 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637497:646298 [6] NCCL INFO comm 0x5555d6c24420 rank 14 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-009:637497:646298 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 cnode7-009:637497:646298 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637495:645505 [4] NCCL INFO comm 0x5555c2c27fe0 rank 12 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-009:637495:645505 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 cnode7-009:637495:645505 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637498:646011 [7] NCCL INFO comm 0x5555d6bd2ca0 rank 15 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-009:637498:646011 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 cnode7-009:637498:646011 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507510:1516380 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-008:1507510:1516380 [5] NCCL INFO comm 0x5555c2bccdd0 rank 5 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507510:1516380 [5] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507510:1516380 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 cnode7-008:1507510:1516380 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637492:646688 [1] NCCL INFO comm 0x5555c2bf3600 rank 9 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-009:637492:646688 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 cnode7-009:637492:646688 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637494:645506 [3] NCCL INFO comm 0x5555c2bfa560 rank 11 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-009:637494:645506 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 cnode7-009:637494:645506 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507509:1516800 [4] NCCL INFO comm 0x5555c2bf6b30 rank 4 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507509:1516800 [4] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507509:1516800 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 cnode7-008:1507509:1516800 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507511:1516731 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-008:1507511:1516731 [6] NCCL INFO comm 0x5555c2bc8c60 rank 6 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 0: 0 8 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 1: 1 9 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 2: 2 10 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 3: 3 11 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 4: 4 12 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 5: 5 13 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 6: 6 14 cnode7-008:1507511:1516731 [6] NCCL INFO NVLS Head 7: 7 15 cnode7-008:1507511:1516731 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 cnode7-008:1507511:1516731 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637493:646472 [2] NCCL INFO comm 0x5555c2c0bbd0 rank 10 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-009:637493:646472 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 cnode7-009:637493:646472 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-009:637491:647053 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 12/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO comm 0x5555c2bf8710 rank 13 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-009:637496:645919 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 cnode7-009:637496:645919 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-008:1507508:1516799 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 08/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 10/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 08/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 14/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Connected all rings cnode7-008:1507506:1516381 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507506:1516381 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Connected all rings cnode7-008:1507509:1516800 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507509:1516800 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Connected all rings cnode7-008:1507511:1516731 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Connected all rings cnode7-008:1507510:1516380 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Connected all rings cnode7-008:1507507:1516100 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-008:1507507:1516100 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Connected all rings cnode7-009:637491:647053 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-009:637491:647053 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Connected all rings cnode7-009:637495:645505 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637495:645505 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Connected all rings cnode7-009:637493:646472 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Connected all rings cnode7-008:1507512:1515700 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-008:1507512:1515700 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Connected all rings cnode7-009:637497:646298 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637497:646298 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Connected all rings cnode7-009:637494:645506 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637494:645506 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Connected all rings cnode7-009:637496:645919 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-009:637496:645919 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Connected all rings cnode7-008:1507508:1516799 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507508:1516799 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Connected all rings cnode7-008:1507505:1515283 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-008:1507505:1515283 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Connected all rings cnode7-009:637492:646688 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637492:646688 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Connected all rings cnode7-009:637498:646011 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-009:637498:646011 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507511:1516731 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-008:1507510:1516380 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-009:637493:646472 [2] NCCL INFO Connected all trees cnode7-008:1507507:1516100 [2] NCCL INFO Connected all trees cnode7-008:1507508:1516799 [3] NCCL INFO Connected all trees cnode7-009:637494:645506 [3] NCCL INFO Connected all trees cnode7-008:1507509:1516800 [4] NCCL INFO Connected all trees cnode7-008:1507511:1516731 [6] NCCL INFO Connected all trees cnode7-008:1507512:1515700 [7] NCCL INFO Connected all trees cnode7-009:637491:647053 [0] NCCL INFO Connected all trees cnode7-009:637497:646298 [6] NCCL INFO Connected all trees cnode7-009:637498:646011 [7] NCCL INFO Connected all trees cnode7-009:637495:645505 [4] NCCL INFO Connected all trees cnode7-009:637496:645919 [5] NCCL INFO Connected all trees cnode7-008:1507510:1516380 [5] NCCL INFO Connected all trees cnode7-008:1507505:1515283 [0] NCCL INFO Connected all trees cnode7-008:1507506:1516381 [1] NCCL INFO Connected all trees cnode7-009:637493:646472 [2] NCCL INFO NVLS comm 0x5555c2c0bbd0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507512:1515700 [7] NCCL INFO NVLS comm 0x5555c2be6e00 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637497:646298 [6] NCCL INFO NVLS comm 0x5555d6c24420 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637494:645506 [3] NCCL INFO NVLS comm 0x5555c2bfa560 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637492:646688 [1] NCCL INFO Connected all trees cnode7-009:637496:645919 [5] NCCL INFO NVLS comm 0x5555c2bf8710 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507508:1516799 [3] NCCL INFO NVLS comm 0x5555c2bd3d50 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507505:1515283 [0] NCCL INFO NVLS comm 0x5555b8bbaf30 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637492:646688 [1] NCCL INFO NVLS comm 0x5555c2bf3600 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637498:646011 [7] NCCL INFO NVLS comm 0x5555d6bd2ca0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507506:1516381 [1] NCCL INFO NVLS comm 0x5555c2c00610 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507506:1516381 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO NVLS comm 0x5555c2bf6b30 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507509:1516800 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507509:1516800 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO NVLS comm 0x5555c2bc8c60 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507511:1516731 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507511:1516731 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO NVLS comm 0x5555c2bccdd0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507510:1516380 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507510:1516380 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO NVLS comm 0x5555c2bf72f0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-008:1507507:1516100 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507507:1516100 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO NVLS comm 0x5555b8bd9f40 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637491:647053 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO NVLS comm 0x5555c2c27fe0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-009:637495:645505 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-009:637493:646472 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-008:1507512:1515700 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637497:646298 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637494:645506 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-009:637496:645919 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507508:1516799 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-008:1507505:1515283 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637492:646688 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637498:646011 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637491:647053 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-009:637495:645505 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-008:1507506:1516381 [1] NCCL INFO Connected NVLS tree cnode7-008:1507506:1516381 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507506:1516381 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637492:646688 [1] NCCL INFO Connected NVLS tree cnode7-009:637492:646688 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637492:646688 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637491:647053 [0] NCCL INFO Connected NVLS tree cnode7-009:637491:647053 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637491:647053 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507508:1516799 [3] NCCL INFO Connected NVLS tree cnode7-008:1507508:1516799 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507508:1516799 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507510:1516380 [5] NCCL INFO Connected NVLS tree cnode7-008:1507510:1516380 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507510:1516380 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507505:1515283 [0] NCCL INFO Connected NVLS tree cnode7-008:1507505:1515283 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507505:1515283 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637493:646472 [2] NCCL INFO Connected NVLS tree cnode7-009:637493:646472 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637493:646472 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637494:645506 [3] NCCL INFO Connected NVLS tree cnode7-009:637494:645506 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637494:645506 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507507:1516100 [2] NCCL INFO Connected NVLS tree cnode7-008:1507507:1516100 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507507:1516100 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507511:1516731 [6] NCCL INFO Connected NVLS tree cnode7-008:1507511:1516731 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507511:1516731 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637496:645919 [5] NCCL INFO Connected NVLS tree cnode7-009:637496:645919 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637496:645919 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507512:1515700 [7] NCCL INFO Connected NVLS tree cnode7-008:1507512:1515700 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507512:1515700 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507509:1516800 [4] NCCL INFO Connected NVLS tree cnode7-008:1507509:1516800 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-008:1507509:1516800 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637495:645505 [4] NCCL INFO Connected NVLS tree cnode7-009:637495:645505 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637495:645505 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637497:646298 [6] NCCL INFO Connected NVLS tree cnode7-009:637497:646298 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637497:646298 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-009:637498:646011 [7] NCCL INFO Connected NVLS tree cnode7-009:637498:646011 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-009:637498:646011 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-008:1507512:1515700 [7] NCCL INFO comm 0x5555c2be6e00 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637496:645919 [5] NCCL INFO comm 0x5555c2bf8710 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507510:1516380 [5] NCCL INFO comm 0x5555c2bccdd0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507511:1516731 [6] NCCL INFO comm 0x5555c2bc8c60 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507509:1516800 [4] NCCL INFO comm 0x5555c2bf6b30 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507508:1516799 [3] NCCL INFO comm 0x5555c2bd3d50 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507506:1516381 [1] NCCL INFO comm 0x5555c2c00610 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507505:1515283 [0] NCCL INFO comm 0x5555b8bbaf30 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-008:1507507:1516100 [2] NCCL INFO comm 0x5555c2bf72f0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637491:647053 [0] NCCL INFO comm 0x5555b8bd9f40 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637494:645506 [3] NCCL INFO comm 0x5555c2bfa560 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637497:646298 [6] NCCL INFO comm 0x5555d6c24420 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637498:646011 [7] NCCL INFO comm 0x5555d6bd2ca0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637495:645505 [4] NCCL INFO comm 0x5555c2c27fe0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637493:646472 [2] NCCL INFO comm 0x5555c2c0bbd0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xebe7a4df1cc421c2 - Init COMPLETE cnode7-009:637492:646688 [1] NCCL INFO comm 0x5555c2bf3600 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xebe7a4df1cc421c2 - Init COMPLETE [cnode7-008:0/16] 2024-11-28 15:22:18,088 (deepspeed_trainer:226) INFO: 1epoch:train:1-100batch: iter_time=2.156, loss_ctc=7.543e+03, loss_att=398.560, acc=2.112e-05, loss=2.542e+03, grad_norm=266.274, loss_scale=1.000, learning_rate=8.583e-08, step_time=0.753 [cnode7-008:0/16] 2024-11-28 15:22:55,224 (deepspeed_trainer:226) INFO: 1epoch:train:101-200batch: iter_time=1.091e-04, loss_ctc=6.971e+03, loss_att=366.335, acc=1.547e-05, loss=2.348e+03, grad_norm=679.559, loss_scale=1.000, learning_rate=2.525e-07, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:23:32,012 (deepspeed_trainer:226) INFO: 1epoch:train:201-300batch: iter_time=1.102e-04, loss_ctc=2.888e+03, loss_att=386.217, acc=5.097e-06, loss=1.137e+03, grad_norm=828.366, loss_scale=1.000, learning_rate=4.192e-07, step_time=0.367 [cnode7-008:0/16] 2024-11-28 15:24:09,182 (deepspeed_trainer:226) INFO: 1epoch:train:301-400batch: iter_time=1.110e-04, loss_ctc=554.115, loss_att=388.485, acc=1.251e-05, loss=438.245, grad_norm=163.601, loss_scale=1.000, learning_rate=5.858e-07, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:24:45,947 (deepspeed_trainer:226) INFO: 1epoch:train:401-500batch: iter_time=1.103e-04, loss_ctc=371.850, loss_att=381.400, acc=4.335e-05, loss=378.335, grad_norm=24.882, loss_scale=1.000, learning_rate=7.525e-07, step_time=0.367 [cnode7-008:0/16] 2024-11-28 15:25:23,002 (deepspeed_trainer:226) INFO: 1epoch:train:501-600batch: iter_time=1.097e-04, loss_ctc=400.150, loss_att=428.135, acc=2.144e-04, loss=419.860, grad_norm=23.466, loss_scale=1.000, learning_rate=9.192e-07, step_time=0.370 [cnode7-008:0/16] 2024-11-28 15:25:59,847 (deepspeed_trainer:226) INFO: 1epoch:train:601-700batch: iter_time=1.082e-04, loss_ctc=349.625, loss_att=375.470, acc=0.002, loss=367.565, grad_norm=17.960, loss_scale=1.000, learning_rate=1.086e-06, step_time=0.368 [cnode7-008:0/16] 2024-11-28 15:26:38,010 (deepspeed_trainer:226) INFO: 1epoch:train:701-800batch: iter_time=1.046e-04, loss_ctc=385.155, loss_att=429.795, acc=0.014, loss=416.555, grad_norm=18.182, loss_scale=1.000, learning_rate=1.252e-06, step_time=0.382 [cnode7-008:0/16] 2024-11-28 15:27:16,384 (deepspeed_trainer:226) INFO: 1epoch:train:801-900batch: iter_time=1.081e-04, loss_ctc=358.965, loss_att=389.020, acc=0.035, loss=380.030, grad_norm=15.344, loss_scale=1.000, learning_rate=1.419e-06, step_time=0.383 [2024-11-28 15:27:53,521] [INFO] [logging.py:129:log_dist] [Rank 0] step=1000, skipped=0, lr=[np.float64(1.6683333333333334e-06)], mom=[[0.9, 0.98]] [2024-11-28 15:27:53,521] [INFO] [timer.py:264:stop] epoch=0/micro_step=1000/global_step=1000, RunningAvgSamplesPerSec=43.72975748596158, CurrSamplesPerSec=44.76266266175005, MemAllocated=1.77GB, MaxMemAllocated=15.24GB [cnode7-008:0/16] 2024-11-28 15:27:53,522 (deepspeed_trainer:226) INFO: 1epoch:train:901-1000batch: iter_time=1.079e-04, loss_ctc=377.145, loss_att=413.100, acc=0.050, loss=402.510, grad_norm=14.377, loss_scale=1.000, learning_rate=1.586e-06, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:28:35,723 (deepspeed_trainer:226) INFO: 1epoch:train:1001-1100batch: iter_time=1.035e-04, loss_ctc=339.174, loss_att=377.034, acc=0.055, loss=365.694, grad_norm=13.825, loss_scale=1.000, learning_rate=1.752e-06, step_time=0.422 [cnode7-008:0/16] 2024-11-28 15:29:13,289 (deepspeed_trainer:226) INFO: 1epoch:train:1101-1200batch: iter_time=1.032e-04, loss_ctc=349.210, loss_att=386.120, acc=0.052, loss=374.970, grad_norm=14.972, loss_scale=1.000, learning_rate=1.919e-06, step_time=0.375 [cnode7-008:0/16] 2024-11-28 15:29:50,626 (deepspeed_trainer:226) INFO: 1epoch:train:1201-1300batch: iter_time=1.012e-04, loss_ctc=344.048, loss_att=365.440, acc=0.052, loss=358.920, grad_norm=13.854, loss_scale=1.000, learning_rate=2.086e-06, step_time=0.373 [cnode7-008:0/16] 2024-11-28 15:30:28,003 (deepspeed_trainer:226) INFO: 1epoch:train:1301-1400batch: iter_time=1.037e-04, loss_ctc=358.790, loss_att=378.215, acc=0.052, loss=372.495, grad_norm=13.807, loss_scale=1.000, learning_rate=2.253e-06, step_time=0.373 [cnode7-008:0/16] 2024-11-28 15:31:05,132 (deepspeed_trainer:226) INFO: 1epoch:train:1401-1500batch: iter_time=1.024e-04, loss_ctc=306.602, loss_att=323.350, acc=0.054, loss=318.380, grad_norm=10.922, loss_scale=1.000, learning_rate=2.419e-06, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:31:42,153 (deepspeed_trainer:226) INFO: 1epoch:train:1501-1600batch: iter_time=1.015e-04, loss_ctc=327.820, loss_att=349.465, acc=0.060, loss=342.990, grad_norm=14.207, loss_scale=1.000, learning_rate=2.586e-06, step_time=0.370 [cnode7-008:0/16] 2024-11-28 15:32:19,094 (deepspeed_trainer:226) INFO: 1epoch:train:1601-1700batch: iter_time=1.027e-04, loss_ctc=322.510, loss_att=336.050, acc=0.075, loss=332.055, grad_norm=12.060, loss_scale=1.000, learning_rate=2.752e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:32:56,476 (deepspeed_trainer:226) INFO: 1epoch:train:1701-1800batch: iter_time=1.040e-04, loss_ctc=352.202, loss_att=359.390, acc=0.087, loss=357.175, grad_norm=11.407, loss_scale=1.000, learning_rate=2.919e-06, step_time=0.373 [cnode7-008:0/16] 2024-11-28 15:33:27,022 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-28 15:33:53,399 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 15:33:59,572 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 15:33:59,572 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-28 15:33:59,575 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 15:34:23,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:25,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:24,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:24,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:26,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:24,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:27,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:27,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:25,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:25,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:25,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:27,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:27,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:28,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:30,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:34:33,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:13,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:11,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:14,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:12,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:12,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:14,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:14,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:12,754] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:15,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:15,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:13,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:15,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:13,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:20,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:21,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:24,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:58,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:00,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:35:59,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:02,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:02,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:00,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:03,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:01,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:07,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:16,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:20,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:47,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:49,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:47,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:47,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:49,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:48,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:51,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:36:54,253] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:37:08,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:37:12,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 15:37:51,644 (deepspeed_trainer:226) INFO: 1epoch:train:1801-1900batch: iter_time=2.304, loss_ctc=309.515, loss_att=321.075, acc=0.103, loss=317.395, grad_norm=10.432, loss_scale=1.000, learning_rate=3.086e-06, step_time=0.648 [2024-11-28 15:38:29,044] [INFO] [logging.py:129:log_dist] [Rank 0] step=2000, skipped=0, lr=[np.float64(3.3350000000000003e-06)], mom=[[0.9, 0.98]] [2024-11-28 15:38:29,044] [INFO] [timer.py:264:stop] epoch=0/micro_step=2000/global_step=2000, RunningAvgSamplesPerSec=41.8745900696839, CurrSamplesPerSec=44.9621291037962, MemAllocated=1.77GB, MaxMemAllocated=15.25GB [cnode7-008:0/16] 2024-11-28 15:38:29,045 (deepspeed_trainer:226) INFO: 1epoch:train:1901-2000batch: iter_time=1.031e-04, loss_ctc=292.725, loss_att=306.245, acc=0.109, loss=302.160, grad_norm=10.115, loss_scale=1.000, learning_rate=3.253e-06, step_time=0.374 [cnode7-008:0/16] 2024-11-28 15:39:06,076 (deepspeed_trainer:226) INFO: 1epoch:train:2001-2100batch: iter_time=1.122e-04, loss_ctc=294.672, loss_att=294.375, acc=0.115, loss=294.385, grad_norm=9.908, loss_scale=1.000, learning_rate=3.419e-06, step_time=0.370 [cnode7-008:0/16] 2024-11-28 15:39:43,202 (deepspeed_trainer:226) INFO: 1epoch:train:2101-2200batch: iter_time=1.093e-04, loss_ctc=277.465, loss_att=296.670, acc=0.118, loss=290.785, grad_norm=8.629, loss_scale=1.000, learning_rate=3.586e-06, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:40:20,659 (deepspeed_trainer:226) INFO: 1epoch:train:2201-2300batch: iter_time=1.091e-04, loss_ctc=318.884, loss_att=324.006, acc=0.108, loss=322.551, grad_norm=8.354, loss_scale=1.000, learning_rate=3.753e-06, step_time=0.374 [cnode7-008:0/16] 2024-11-28 15:40:57,519 (deepspeed_trainer:226) INFO: 1epoch:train:2301-2400batch: iter_time=1.102e-04, loss_ctc=269.223, loss_att=269.885, acc=0.125, loss=269.810, grad_norm=8.459, loss_scale=1.000, learning_rate=3.919e-06, step_time=0.368 [cnode7-008:0/16] 2024-11-28 15:41:34,500 (deepspeed_trainer:226) INFO: 1epoch:train:2401-2500batch: iter_time=1.111e-04, loss_ctc=317.727, loss_att=322.579, acc=0.113, loss=321.104, grad_norm=7.948, loss_scale=1.000, learning_rate=4.086e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:42:11,395 (deepspeed_trainer:226) INFO: 1epoch:train:2501-2600batch: iter_time=1.103e-04, loss_ctc=302.751, loss_att=307.864, acc=0.117, loss=306.349, grad_norm=7.740, loss_scale=1.000, learning_rate=4.253e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:42:48,324 (deepspeed_trainer:226) INFO: 1epoch:train:2601-2700batch: iter_time=1.115e-04, loss_ctc=295.397, loss_att=308.217, acc=0.120, loss=304.335, grad_norm=7.659, loss_scale=1.000, learning_rate=4.419e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:43:25,265 (deepspeed_trainer:226) INFO: 1epoch:train:2701-2800batch: iter_time=1.114e-04, loss_ctc=290.483, loss_att=298.595, acc=0.121, loss=296.192, grad_norm=7.037, loss_scale=1.000, learning_rate=4.586e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:44:02,130 (deepspeed_trainer:226) INFO: 1epoch:train:2801-2900batch: iter_time=1.079e-04, loss_ctc=299.820, loss_att=302.995, acc=0.125, loss=302.017, grad_norm=7.540, loss_scale=1.000, learning_rate=4.752e-06, step_time=0.368 [2024-11-28 15:44:40,073] [INFO] [logging.py:129:log_dist] [Rank 0] step=3000, skipped=0, lr=[np.float64(5.0016666666666665e-06)], mom=[[0.9, 0.98]] [2024-11-28 15:44:40,074] [INFO] [timer.py:264:stop] epoch=0/micro_step=3000/global_step=3000, RunningAvgSamplesPerSec=42.55412757157384, CurrSamplesPerSec=43.68466018788812, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 15:44:40,075 (deepspeed_trainer:226) INFO: 1epoch:train:2901-3000batch: iter_time=1.034e-04, loss_ctc=285.794, loss_att=302.037, acc=0.133, loss=297.210, grad_norm=7.693, loss_scale=1.000, learning_rate=4.919e-06, step_time=0.379 [cnode7-008:0/16] 2024-11-28 15:45:17,484 (deepspeed_trainer:226) INFO: 1epoch:train:3001-3100batch: iter_time=1.061e-04, loss_ctc=284.330, loss_att=291.895, acc=0.128, loss=289.658, grad_norm=7.779, loss_scale=1.000, learning_rate=5.086e-06, step_time=0.374 [cnode7-008:0/16] 2024-11-28 15:45:58,113 (deepspeed_trainer:226) INFO: 1epoch:train:3101-3200batch: iter_time=1.078e-04, loss_ctc=293.512, loss_att=293.335, acc=0.126, loss=293.337, grad_norm=7.667, loss_scale=1.000, learning_rate=5.252e-06, step_time=0.406 [cnode7-008:0/16] 2024-11-28 15:46:35,886 (deepspeed_trainer:226) INFO: 1epoch:train:3201-3300batch: iter_time=1.081e-04, loss_ctc=283.230, loss_att=293.402, acc=0.131, loss=290.353, grad_norm=8.681, loss_scale=1.000, learning_rate=5.419e-06, step_time=0.377 [cnode7-008:0/16] 2024-11-28 15:47:12,597 (deepspeed_trainer:226) INFO: 1epoch:train:3301-3400batch: iter_time=1.096e-04, loss_ctc=281.617, loss_att=279.262, acc=0.133, loss=279.897, grad_norm=7.327, loss_scale=1.000, learning_rate=5.586e-06, step_time=0.367 [cnode7-008:0/16] 2024-11-28 15:47:49,589 (deepspeed_trainer:226) INFO: 1epoch:train:3401-3500batch: iter_time=1.082e-04, loss_ctc=271.657, loss_att=283.540, acc=0.138, loss=279.874, grad_norm=8.653, loss_scale=1.000, learning_rate=5.753e-06, step_time=0.370 [cnode7-008:0/16] 2024-11-28 15:48:26,680 (deepspeed_trainer:226) INFO: 1epoch:train:3501-3600batch: iter_time=1.094e-04, loss_ctc=297.170, loss_att=296.858, acc=0.133, loss=297.058, grad_norm=10.215, loss_scale=1.000, learning_rate=5.919e-06, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:49:03,830 (deepspeed_trainer:226) INFO: 1epoch:train:3601-3700batch: iter_time=1.087e-04, loss_ctc=302.332, loss_att=301.413, acc=0.139, loss=301.680, grad_norm=8.118, loss_scale=1.000, learning_rate=6.086e-06, step_time=0.371 [cnode7-008:0/16] 2024-11-28 15:49:25,389 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-28 15:49:51,871 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 15:49:57,416 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 15:49:57,417 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-28 15:49:57,419 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 15:50:21,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:21,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:24,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:24,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:22,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:22,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:22,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:25,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:23,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:23,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:25,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:25,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:26,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:23,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:26,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:50:26,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:09,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:12,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:10,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:12,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:10,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:11,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:11,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:11,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:11,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:14,477] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:12,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:56,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:57,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:00,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:00,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:58,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:58,968] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:51:59,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:01,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:01,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:02,068] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:02,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:00,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:02,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:02,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:00,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:00,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:44,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:46,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:44,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:47,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:49,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:46,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:46,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:49,473] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:47,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:49,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:47,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:50,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:47,824] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:50,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:48,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 15:52:52,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 15:53:36,801 (deepspeed_trainer:226) INFO: 1epoch:train:3701-3800batch: iter_time=2.328, loss_ctc=265.850, loss_att=276.582, acc=0.146, loss=273.423, grad_norm=9.590, loss_scale=1.000, learning_rate=6.253e-06, step_time=0.401 [cnode7-008:0/16] 2024-11-28 15:54:15,091 (deepspeed_trainer:226) INFO: 1epoch:train:3801-3900batch: iter_time=1.050e-04, loss_ctc=278.785, loss_att=295.663, acc=0.145, loss=290.502, grad_norm=9.275, loss_scale=1.000, learning_rate=6.419e-06, step_time=0.383 [2024-11-28 15:54:53,193] [INFO] [logging.py:129:log_dist] [Rank 0] step=4000, skipped=0, lr=[np.float64(6.668333333333334e-06)], mom=[[0.9, 0.98]] [2024-11-28 15:54:53,193] [INFO] [timer.py:264:stop] epoch=0/micro_step=4000/global_step=4000, RunningAvgSamplesPerSec=42.636254678561855, CurrSamplesPerSec=43.483322718377, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 15:54:53,195 (deepspeed_trainer:226) INFO: 1epoch:train:3901-4000batch: iter_time=1.049e-04, loss_ctc=271.187, loss_att=286.807, acc=0.144, loss=282.017, grad_norm=8.283, loss_scale=1.000, learning_rate=6.586e-06, step_time=0.381 [cnode7-008:0/16] 2024-11-28 15:55:34,003 (deepspeed_trainer:226) INFO: 1epoch:train:4001-4100batch: iter_time=1.056e-04, loss_ctc=272.600, loss_att=293.283, acc=0.147, loss=287.053, grad_norm=9.053, loss_scale=1.000, learning_rate=6.752e-06, step_time=0.407 [cnode7-008:0/16] 2024-11-28 15:56:12,637 (deepspeed_trainer:226) INFO: 1epoch:train:4101-4200batch: iter_time=1.072e-04, loss_ctc=276.420, loss_att=291.010, acc=0.146, loss=286.637, grad_norm=11.222, loss_scale=1.000, learning_rate=6.919e-06, step_time=0.386 [cnode7-008:0/16] 2024-11-28 15:56:50,417 (deepspeed_trainer:226) INFO: 1epoch:train:4201-4300batch: iter_time=1.058e-04, loss_ctc=285.565, loss_att=294.673, acc=0.149, loss=291.908, grad_norm=9.771, loss_scale=1.000, learning_rate=7.086e-06, step_time=0.377 [cnode7-008:0/16] 2024-11-28 15:57:28,147 (deepspeed_trainer:226) INFO: 1epoch:train:4301-4400batch: iter_time=1.059e-04, loss_ctc=306.363, loss_att=313.845, acc=0.140, loss=311.600, grad_norm=9.831, loss_scale=1.000, learning_rate=7.253e-06, step_time=0.377 [cnode7-008:0/16] 2024-11-28 15:58:05,487 (deepspeed_trainer:226) INFO: 1epoch:train:4401-4500batch: iter_time=1.068e-04, loss_ctc=277.127, loss_att=286.950, acc=0.147, loss=283.873, grad_norm=9.893, loss_scale=1.000, learning_rate=7.419e-06, step_time=0.373 [cnode7-008:0/16] 2024-11-28 15:58:42,537 (deepspeed_trainer:226) INFO: 1epoch:train:4501-4600batch: iter_time=1.053e-04, loss_ctc=280.988, loss_att=295.408, acc=0.149, loss=291.108, grad_norm=8.921, loss_scale=1.000, learning_rate=7.586e-06, step_time=0.370 [cnode7-008:0/16] 2024-11-28 15:59:19,441 (deepspeed_trainer:226) INFO: 1epoch:train:4601-4700batch: iter_time=1.077e-04, loss_ctc=278.002, loss_att=280.447, acc=0.146, loss=279.762, grad_norm=8.641, loss_scale=1.000, learning_rate=7.753e-06, step_time=0.369 [cnode7-008:0/16] 2024-11-28 15:59:56,683 (deepspeed_trainer:226) INFO: 1epoch:train:4701-4800batch: iter_time=1.084e-04, loss_ctc=300.358, loss_att=297.410, acc=0.150, loss=298.233, grad_norm=8.880, loss_scale=1.000, learning_rate=7.919e-06, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:00:34,572 (deepspeed_trainer:226) INFO: 1epoch:train:4801-4900batch: iter_time=1.088e-04, loss_ctc=290.676, loss_att=304.460, acc=0.147, loss=300.348, grad_norm=9.107, loss_scale=1.000, learning_rate=8.086e-06, step_time=0.379 [2024-11-28 16:01:12,193] [INFO] [logging.py:129:log_dist] [Rank 0] step=5000, skipped=0, lr=[np.float64(8.335e-06)], mom=[[0.9, 0.98]] [2024-11-28 16:01:12,194] [INFO] [timer.py:264:stop] epoch=0/micro_step=5000/global_step=5000, RunningAvgSamplesPerSec=42.717182720769884, CurrSamplesPerSec=41.498636437455986, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:01:12,195 (deepspeed_trainer:226) INFO: 1epoch:train:4901-5000batch: iter_time=1.098e-04, loss_ctc=276.480, loss_att=285.245, acc=0.149, loss=282.550, grad_norm=13.674, loss_scale=1.000, learning_rate=8.252e-06, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:01:49,971 (deepspeed_trainer:226) INFO: 1epoch:train:5001-5100batch: iter_time=1.082e-04, loss_ctc=279.072, loss_att=283.980, acc=0.149, loss=282.475, grad_norm=9.322, loss_scale=1.000, learning_rate=8.419e-06, step_time=0.378 [cnode7-008:0/16] 2024-11-28 16:02:28,158 (deepspeed_trainer:226) INFO: 1epoch:train:5101-5200batch: iter_time=1.050e-04, loss_ctc=287.812, loss_att=301.532, acc=0.153, loss=297.415, grad_norm=7.546, loss_scale=1.000, learning_rate=8.586e-06, step_time=0.381 [cnode7-008:0/16] 2024-11-28 16:03:05,437 (deepspeed_trainer:226) INFO: 1epoch:train:5201-5300batch: iter_time=1.123e-04, loss_ctc=249.620, loss_att=244.730, acc=0.161, loss=246.167, grad_norm=7.901, loss_scale=1.000, learning_rate=8.753e-06, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:03:46,756 (deepspeed_trainer:226) INFO: 1epoch:train:5301-5400batch: iter_time=1.093e-04, loss_ctc=295.702, loss_att=304.950, acc=0.156, loss=302.137, grad_norm=9.180, loss_scale=1.000, learning_rate=8.919e-06, step_time=0.413 [cnode7-008:0/16] 2024-11-28 16:04:26,073 (deepspeed_trainer:226) INFO: 1epoch:train:5401-5500batch: iter_time=1.102e-04, loss_ctc=291.777, loss_att=291.295, acc=0.154, loss=291.375, grad_norm=9.645, loss_scale=1.000, learning_rate=9.086e-06, step_time=0.392 [cnode7-008:0/16] 2024-11-28 16:05:03,580 (deepspeed_trainer:226) INFO: 1epoch:train:5501-5600batch: iter_time=1.091e-04, loss_ctc=299.082, loss_att=294.860, acc=0.155, loss=296.108, grad_norm=9.718, loss_scale=1.000, learning_rate=9.252e-06, step_time=0.375 [cnode7-008:0/16] 2024-11-28 16:05:16,042 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-28 16:05:42,868 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 16:05:48,445 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 16:05:48,445 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-28 16:05:48,448 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 16:06:13,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:15,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:15,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:13,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:13,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:16,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:14,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:14,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:14,540] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:16,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:17,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:17,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:17,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:17,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:15,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:06:15,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:00,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:01,586] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:04,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:02,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:05,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:05,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:05,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,495] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:03,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:08,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:07,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:46,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:49,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:50,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:50,859] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:51,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:49,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:52,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:52,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:50,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:50,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:53,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:51,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:51,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:51,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:58,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:07:57,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:34,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:37,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:37,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:39,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:37,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:37,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:41,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:38,900] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:41,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:39,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:39,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:40,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:43,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:46,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:48,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:08:47,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 16:09:43,319 (deepspeed_trainer:226) INFO: 1epoch:train:5601-5700batch: iter_time=2.374, loss_ctc=270.270, loss_att=270.395, acc=0.164, loss=270.420, grad_norm=8.470, loss_scale=1.000, learning_rate=9.419e-06, step_time=0.423 [cnode7-008:0/16] 2024-11-28 16:10:20,994 (deepspeed_trainer:226) INFO: 1epoch:train:5701-5800batch: iter_time=1.076e-04, loss_ctc=268.855, loss_att=260.845, acc=0.165, loss=263.295, grad_norm=9.334, loss_scale=1.000, learning_rate=9.586e-06, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:10:58,328 (deepspeed_trainer:226) INFO: 1epoch:train:5801-5900batch: iter_time=1.086e-04, loss_ctc=260.817, loss_att=251.345, acc=0.174, loss=254.077, grad_norm=8.661, loss_scale=1.000, learning_rate=9.753e-06, step_time=0.373 [2024-11-28 16:11:35,828] [INFO] [logging.py:129:log_dist] [Rank 0] step=6000, skipped=0, lr=[np.float64(1.0001666666666666e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:11:35,829] [INFO] [timer.py:264:stop] epoch=0/micro_step=6000/global_step=6000, RunningAvgSamplesPerSec=42.63483416617961, CurrSamplesPerSec=44.40845314948052, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:11:35,830 (deepspeed_trainer:226) INFO: 1epoch:train:5901-6000batch: iter_time=1.085e-04, loss_ctc=266.043, loss_att=261.137, acc=0.175, loss=262.610, grad_norm=10.708, loss_scale=1.000, learning_rate=9.919e-06, step_time=0.375 [cnode7-008:0/16] 2024-11-28 16:12:13,561 (deepspeed_trainer:226) INFO: 1epoch:train:6001-6100batch: iter_time=1.021e-04, loss_ctc=272.558, loss_att=257.290, acc=0.165, loss=261.783, grad_norm=13.046, loss_scale=1.000, learning_rate=1.009e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 16:12:51,299 (deepspeed_trainer:226) INFO: 1epoch:train:6101-6200batch: iter_time=1.055e-04, loss_ctc=292.705, loss_att=275.130, acc=0.173, loss=280.445, grad_norm=8.975, loss_scale=1.000, learning_rate=1.025e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 16:13:29,485 (deepspeed_trainer:226) INFO: 1epoch:train:6201-6300batch: iter_time=1.083e-04, loss_ctc=285.363, loss_att=265.517, acc=0.165, loss=271.475, grad_norm=8.770, loss_scale=1.000, learning_rate=1.042e-05, step_time=0.382 [cnode7-008:0/16] 2024-11-28 16:14:11,733 (deepspeed_trainer:226) INFO: 1epoch:train:6301-6400batch: iter_time=1.081e-04, loss_ctc=297.890, loss_att=289.355, acc=0.167, loss=291.900, grad_norm=10.114, loss_scale=1.000, learning_rate=1.059e-05, step_time=0.422 [cnode7-008:0/16] 2024-11-28 16:14:48,900 (deepspeed_trainer:226) INFO: 1epoch:train:6401-6500batch: iter_time=1.088e-04, loss_ctc=264.627, loss_att=253.463, acc=0.170, loss=256.855, grad_norm=7.899, loss_scale=1.000, learning_rate=1.075e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 16:15:26,184 (deepspeed_trainer:226) INFO: 1epoch:train:6501-6600batch: iter_time=1.070e-04, loss_ctc=302.808, loss_att=280.627, acc=0.165, loss=287.195, grad_norm=9.529, loss_scale=1.000, learning_rate=1.092e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:16:03,072 (deepspeed_trainer:226) INFO: 1epoch:train:6601-6700batch: iter_time=1.067e-04, loss_ctc=264.090, loss_att=243.548, acc=0.180, loss=249.657, grad_norm=10.714, loss_scale=1.000, learning_rate=1.109e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 16:16:40,286 (deepspeed_trainer:226) INFO: 1epoch:train:6701-6800batch: iter_time=1.059e-04, loss_ctc=277.439, loss_att=272.252, acc=0.170, loss=273.800, grad_norm=8.304, loss_scale=1.000, learning_rate=1.125e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:17:17,564 (deepspeed_trainer:226) INFO: 1epoch:train:6801-6900batch: iter_time=1.084e-04, loss_ctc=292.212, loss_att=269.142, acc=0.173, loss=276.070, grad_norm=10.665, loss_scale=1.000, learning_rate=1.142e-05, step_time=0.372 [2024-11-28 16:17:54,632] [INFO] [logging.py:129:log_dist] [Rank 0] step=7000, skipped=0, lr=[np.float64(1.1668333333333333e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:17:54,633] [INFO] [timer.py:264:stop] epoch=0/micro_step=7000/global_step=7000, RunningAvgSamplesPerSec=42.694925496772605, CurrSamplesPerSec=43.88448425406821, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:17:54,634 (deepspeed_trainer:226) INFO: 1epoch:train:6901-7000batch: iter_time=1.077e-04, loss_ctc=276.572, loss_att=256.382, acc=0.175, loss=262.460, grad_norm=8.304, loss_scale=1.000, learning_rate=1.159e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 16:18:31,987 (deepspeed_trainer:226) INFO: 1epoch:train:7001-7100batch: iter_time=1.086e-04, loss_ctc=265.310, loss_att=253.831, acc=0.177, loss=257.341, grad_norm=8.065, loss_scale=1.000, learning_rate=1.175e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:19:09,016 (deepspeed_trainer:226) INFO: 1epoch:train:7101-7200batch: iter_time=1.087e-04, loss_ctc=277.600, loss_att=256.423, acc=0.181, loss=262.730, grad_norm=8.313, loss_scale=1.000, learning_rate=1.192e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 16:19:46,189 (deepspeed_trainer:226) INFO: 1epoch:train:7201-7300batch: iter_time=1.064e-04, loss_ctc=282.055, loss_att=261.062, acc=0.182, loss=267.442, grad_norm=7.475, loss_scale=1.000, learning_rate=1.209e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:20:23,860 (deepspeed_trainer:226) INFO: 1epoch:train:7301-7400batch: iter_time=1.093e-04, loss_ctc=287.935, loss_att=262.413, acc=0.182, loss=270.090, grad_norm=9.591, loss_scale=1.000, learning_rate=1.225e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:21:01,632 (deepspeed_trainer:226) INFO: 1epoch:train:7401-7500batch: iter_time=1.065e-04, loss_ctc=284.235, loss_att=257.740, acc=0.185, loss=265.605, grad_norm=10.492, loss_scale=1.000, learning_rate=1.242e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 16:21:04,930 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-28 16:21:31,121 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 16:21:36,580 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 16:21:36,580 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-28 16:21:36,583 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 16:22:02,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:04,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:02,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:04,824] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:04,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:04,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:02,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:02,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:05,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,783] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:03,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:04,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:51,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,213] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:49,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,643] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,920] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:50,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:50,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:51,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:51,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:52,354] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:55,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:55,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:22:55,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:38,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:36,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:39,693] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:39,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:37,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:40,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:40,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:40,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:38,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:39,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:39,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:39,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:40,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:44,730] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:45,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:23:45,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:25,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:26,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:26,704] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:24,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:24,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:27,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:28,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:28,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:26,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:26,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:27,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:27,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:27,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:33,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:33,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:24:35,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 16:25:40,444 (deepspeed_trainer:226) INFO: 1epoch:train:7501-7600batch: iter_time=2.323, loss_ctc=272.075, loss_att=267.385, acc=0.178, loss=268.790, grad_norm=9.696, loss_scale=1.000, learning_rate=1.259e-05, step_time=0.465 [cnode7-008:0/16] 2024-11-28 16:26:17,792 (deepspeed_trainer:226) INFO: 1epoch:train:7601-7700batch: iter_time=1.087e-04, loss_ctc=260.500, loss_att=246.765, acc=0.182, loss=250.833, grad_norm=8.271, loss_scale=1.000, learning_rate=1.275e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:26:55,459 (deepspeed_trainer:226) INFO: 1epoch:train:7701-7800batch: iter_time=1.096e-04, loss_ctc=256.597, loss_att=250.649, acc=0.195, loss=252.467, grad_norm=9.820, loss_scale=1.000, learning_rate=1.292e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:27:33,167 (deepspeed_trainer:226) INFO: 1epoch:train:7801-7900batch: iter_time=1.104e-04, loss_ctc=260.210, loss_att=251.705, acc=0.186, loss=254.110, grad_norm=8.381, loss_scale=1.000, learning_rate=1.309e-05, step_time=0.377 [2024-11-28 16:28:10,542] [INFO] [logging.py:129:log_dist] [Rank 0] step=8000, skipped=0, lr=[np.float64(1.3335e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:28:10,542] [INFO] [timer.py:264:stop] epoch=0/micro_step=8000/global_step=8000, RunningAvgSamplesPerSec=42.67186027793642, CurrSamplesPerSec=40.47910813839216, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:28:10,544 (deepspeed_trainer:226) INFO: 1epoch:train:7901-8000batch: iter_time=1.108e-04, loss_ctc=271.300, loss_att=256.332, acc=0.186, loss=260.780, grad_norm=9.334, loss_scale=1.000, learning_rate=1.325e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:28:47,724 (deepspeed_trainer:226) INFO: 1epoch:train:8001-8100batch: iter_time=1.103e-04, loss_ctc=300.092, loss_att=279.415, acc=0.184, loss=285.697, grad_norm=8.484, loss_scale=1.000, learning_rate=1.342e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:29:24,596 (deepspeed_trainer:226) INFO: 1epoch:train:8101-8200batch: iter_time=1.114e-04, loss_ctc=268.225, loss_att=241.202, acc=0.182, loss=249.282, grad_norm=9.641, loss_scale=1.000, learning_rate=1.359e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 16:30:02,088 (deepspeed_trainer:226) INFO: 1epoch:train:8201-8300batch: iter_time=1.117e-04, loss_ctc=291.123, loss_att=284.145, acc=0.184, loss=286.202, grad_norm=9.238, loss_scale=1.000, learning_rate=1.375e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:30:39,493 (deepspeed_trainer:226) INFO: 1epoch:train:8301-8400batch: iter_time=1.084e-04, loss_ctc=278.265, loss_att=252.572, acc=0.182, loss=260.295, grad_norm=9.914, loss_scale=1.000, learning_rate=1.392e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:31:17,407 (deepspeed_trainer:226) INFO: 1epoch:train:8401-8500batch: iter_time=1.109e-04, loss_ctc=296.553, loss_att=271.152, acc=0.185, loss=278.762, grad_norm=10.166, loss_scale=1.000, learning_rate=1.409e-05, step_time=0.379 [cnode7-008:0/16] 2024-11-28 16:31:54,449 (deepspeed_trainer:226) INFO: 1epoch:train:8501-8600batch: iter_time=1.118e-04, loss_ctc=264.778, loss_att=246.353, acc=0.196, loss=251.961, grad_norm=7.707, loss_scale=1.000, learning_rate=1.425e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 16:32:33,434 (deepspeed_trainer:226) INFO: 1epoch:train:8601-8700batch: iter_time=1.132e-04, loss_ctc=278.978, loss_att=262.717, acc=0.182, loss=267.735, grad_norm=9.377, loss_scale=1.000, learning_rate=1.442e-05, step_time=0.389 [cnode7-008:0/16] 2024-11-28 16:33:13,356 (deepspeed_trainer:226) INFO: 1epoch:train:8701-8800batch: iter_time=1.105e-04, loss_ctc=279.515, loss_att=256.252, acc=0.183, loss=263.210, grad_norm=12.593, loss_scale=1.000, learning_rate=1.459e-05, step_time=0.399 [cnode7-008:0/16] 2024-11-28 16:33:51,137 (deepspeed_trainer:226) INFO: 1epoch:train:8801-8900batch: iter_time=1.113e-04, loss_ctc=294.717, loss_att=277.970, acc=0.185, loss=283.055, grad_norm=9.978, loss_scale=1.000, learning_rate=1.475e-05, step_time=0.377 [2024-11-28 16:34:27,895] [INFO] [logging.py:129:log_dist] [Rank 0] step=9000, skipped=0, lr=[np.float64(1.5001666666666667e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:34:27,896] [INFO] [timer.py:264:stop] epoch=0/micro_step=9000/global_step=9000, RunningAvgSamplesPerSec=42.731397143170746, CurrSamplesPerSec=41.653025749873066, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:34:27,897 (deepspeed_trainer:226) INFO: 1epoch:train:8901-9000batch: iter_time=1.118e-04, loss_ctc=252.468, loss_att=231.483, acc=0.192, loss=237.751, grad_norm=7.587, loss_scale=1.000, learning_rate=1.492e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 16:35:04,773 (deepspeed_trainer:226) INFO: 1epoch:train:9001-9100batch: iter_time=1.096e-04, loss_ctc=279.305, loss_att=256.317, acc=0.193, loss=263.210, grad_norm=9.210, loss_scale=1.000, learning_rate=1.509e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 16:35:41,639 (deepspeed_trainer:226) INFO: 1epoch:train:9101-9200batch: iter_time=1.084e-04, loss_ctc=272.743, loss_att=247.243, acc=0.188, loss=254.870, grad_norm=9.214, loss_scale=1.000, learning_rate=1.525e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 16:36:18,654 (deepspeed_trainer:226) INFO: 1epoch:train:9201-9300batch: iter_time=1.086e-04, loss_ctc=296.548, loss_att=275.772, acc=0.190, loss=281.938, grad_norm=10.548, loss_scale=1.000, learning_rate=1.542e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 16:36:50,013 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-28 16:37:16,137 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 16:37:21,759 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 16:37:21,759 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-28 16:37:21,761 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 16:37:49,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:47,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:49,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:47,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:47,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:50,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:50,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:50,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:48,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:51,448] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:48,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:51,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:51,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:49,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:49,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:37:49,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:37,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:38,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:38,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:36,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:36,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:39,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:39,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:37,184] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:37,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:40,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:37,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:38,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:38,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:38,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:42,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:38:48,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:25,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:25,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:26,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:26,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:25,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:27,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:27,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:25,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:25,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:26,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:27,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:32,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:30,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:30,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:36,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:39:42,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:13,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:13,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:13,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:14,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:15,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:15,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:13,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:13,478] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:14,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:14,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:14,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:22,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:20,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:22,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:29,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:40:34,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 16:41:08,174 (deepspeed_trainer:226) INFO: 1epoch:train:9301-9400batch: iter_time=2.333, loss_ctc=268.507, loss_att=243.273, acc=0.193, loss=250.885, grad_norm=12.027, loss_scale=1.000, learning_rate=1.559e-05, step_time=0.562 [cnode7-008:0/16] 2024-11-28 16:41:46,250 (deepspeed_trainer:226) INFO: 1epoch:train:9401-9500batch: iter_time=1.022e-04, loss_ctc=260.440, loss_att=237.725, acc=0.199, loss=244.572, grad_norm=8.396, loss_scale=1.000, learning_rate=1.575e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 16:42:23,464 (deepspeed_trainer:226) INFO: 1epoch:train:9501-9600batch: iter_time=1.069e-04, loss_ctc=266.650, loss_att=235.165, acc=0.190, loss=244.620, grad_norm=10.707, loss_scale=1.000, learning_rate=1.592e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:43:02,833 (deepspeed_trainer:226) INFO: 1epoch:train:9601-9700batch: iter_time=1.070e-04, loss_ctc=248.045, loss_att=235.160, acc=0.213, loss=238.988, grad_norm=12.665, loss_scale=1.000, learning_rate=1.609e-05, step_time=0.393 [cnode7-008:0/16] 2024-11-28 16:43:42,973 (deepspeed_trainer:226) INFO: 1epoch:train:9701-9800batch: iter_time=1.127e-04, loss_ctc=289.035, loss_att=262.678, acc=0.189, loss=270.610, grad_norm=10.115, loss_scale=1.000, learning_rate=1.625e-05, step_time=0.401 [cnode7-008:0/16] 2024-11-28 16:44:20,609 (deepspeed_trainer:226) INFO: 1epoch:train:9801-9900batch: iter_time=1.076e-04, loss_ctc=251.627, loss_att=222.583, acc=0.201, loss=231.343, grad_norm=9.656, loss_scale=1.000, learning_rate=1.642e-05, step_time=0.376 [2024-11-28 16:44:58,264] [INFO] [logging.py:129:log_dist] [Rank 0] step=10000, skipped=0, lr=[np.float64(1.6668333333333334e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:44:58,265] [INFO] [timer.py:264:stop] epoch=0/micro_step=10000/global_step=10000, RunningAvgSamplesPerSec=42.556442631061365, CurrSamplesPerSec=44.79529070449879, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:44:58,266 (deepspeed_trainer:226) INFO: 1epoch:train:9901-10000batch: iter_time=1.099e-04, loss_ctc=295.587, loss_att=262.092, acc=0.191, loss=272.015, grad_norm=9.378, loss_scale=1.000, learning_rate=1.659e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:45:35,900 (deepspeed_trainer:226) INFO: 1epoch:train:10001-10100batch: iter_time=1.071e-04, loss_ctc=282.710, loss_att=249.722, acc=0.188, loss=259.623, grad_norm=9.797, loss_scale=1.000, learning_rate=1.675e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 16:46:13,376 (deepspeed_trainer:226) INFO: 1epoch:train:10101-10200batch: iter_time=1.079e-04, loss_ctc=279.723, loss_att=254.705, acc=0.194, loss=262.207, grad_norm=14.782, loss_scale=1.000, learning_rate=1.692e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:46:50,663 (deepspeed_trainer:226) INFO: 1epoch:train:10201-10300batch: iter_time=1.081e-04, loss_ctc=276.327, loss_att=241.120, acc=0.190, loss=251.725, grad_norm=9.986, loss_scale=1.000, learning_rate=1.709e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:47:28,028 (deepspeed_trainer:226) INFO: 1epoch:train:10301-10400batch: iter_time=1.075e-04, loss_ctc=285.830, loss_att=248.000, acc=0.198, loss=259.400, grad_norm=9.742, loss_scale=1.000, learning_rate=1.725e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:48:05,363 (deepspeed_trainer:226) INFO: 1epoch:train:10401-10500batch: iter_time=1.083e-04, loss_ctc=272.971, loss_att=250.265, acc=0.199, loss=257.051, grad_norm=8.556, loss_scale=1.000, learning_rate=1.742e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:48:42,749 (deepspeed_trainer:226) INFO: 1epoch:train:10501-10600batch: iter_time=1.075e-04, loss_ctc=270.315, loss_att=243.303, acc=0.194, loss=251.490, grad_norm=8.272, loss_scale=1.000, learning_rate=1.759e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:49:20,073 (deepspeed_trainer:226) INFO: 1epoch:train:10601-10700batch: iter_time=1.069e-04, loss_ctc=282.192, loss_att=246.775, acc=0.192, loss=257.370, grad_norm=11.761, loss_scale=1.000, learning_rate=1.775e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:49:57,490 (deepspeed_trainer:226) INFO: 1epoch:train:10701-10800batch: iter_time=1.085e-04, loss_ctc=271.120, loss_att=246.583, acc=0.198, loss=253.923, grad_norm=10.119, loss_scale=1.000, learning_rate=1.792e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:50:36,285 (deepspeed_trainer:226) INFO: 1epoch:train:10801-10900batch: iter_time=1.058e-04, loss_ctc=271.615, loss_att=235.113, acc=0.197, loss=246.108, grad_norm=9.423, loss_scale=1.000, learning_rate=1.809e-05, step_time=0.388 [2024-11-28 16:51:13,732] [INFO] [logging.py:129:log_dist] [Rank 0] step=11000, skipped=0, lr=[np.float64(1.8335e-05)], mom=[[0.9, 0.98]] [2024-11-28 16:51:13,732] [INFO] [timer.py:264:stop] epoch=0/micro_step=11000/global_step=11000, RunningAvgSamplesPerSec=42.63762946981071, CurrSamplesPerSec=42.862872028463656, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 16:51:13,734 (deepspeed_trainer:226) INFO: 1epoch:train:10901-11000batch: iter_time=1.086e-04, loss_ctc=261.587, loss_att=233.252, acc=0.205, loss=241.725, grad_norm=9.107, loss_scale=1.000, learning_rate=1.825e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 16:51:53,377 (deepspeed_trainer:226) INFO: 1epoch:train:11001-11100batch: iter_time=1.082e-04, loss_ctc=286.550, loss_att=251.325, acc=0.196, loss=261.882, grad_norm=12.946, loss_scale=1.000, learning_rate=1.842e-05, step_time=0.396 [cnode7-008:0/16] 2024-11-28 16:52:33,587 (deepspeed_trainer:226) INFO: 1epoch:train:11101-11200batch: iter_time=1.078e-04, loss_ctc=290.613, loss_att=259.650, acc=0.197, loss=268.967, grad_norm=9.763, loss_scale=1.000, learning_rate=1.859e-05, step_time=0.401 [cnode7-008:0/16] 2024-11-28 16:52:56,018 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-28 16:53:22,210 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 16:53:27,713 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 16:53:27,713 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-28 16:53:27,715 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 16:53:54,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:52,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:55,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:52,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:55,655] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:55,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:55,774] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:56,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:53,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:53,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:53,512] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:53,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:56,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:53,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:56,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:53:57,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:43,249] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:40,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:43,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:40,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:43,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:43,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:41,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:41,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:44,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:42,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:54:48,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:30,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:30,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:27,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:31,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:31,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:32,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:29,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:30,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:30,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:33,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:55:38,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:16,627] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:16,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:17,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:17,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:18,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:18,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:16,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:19,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:19,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:17,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:17,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:17,790] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:18,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:22,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:23,839] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 16:56:28,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 16:57:14,602 (deepspeed_trainer:226) INFO: 1epoch:train:11201-11300batch: iter_time=2.331, loss_ctc=259.010, loss_att=231.070, acc=0.199, loss=239.420, grad_norm=9.311, loss_scale=1.000, learning_rate=1.875e-05, step_time=0.479 [cnode7-008:0/16] 2024-11-28 16:57:51,933 (deepspeed_trainer:226) INFO: 1epoch:train:11301-11400batch: iter_time=1.059e-04, loss_ctc=267.850, loss_att=228.053, acc=0.204, loss=240.010, grad_norm=10.144, loss_scale=1.000, learning_rate=1.892e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 16:58:29,155 (deepspeed_trainer:226) INFO: 1epoch:train:11401-11500batch: iter_time=1.077e-04, loss_ctc=257.847, loss_att=220.988, acc=0.208, loss=232.042, grad_norm=7.842, loss_scale=1.000, learning_rate=1.909e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 16:59:06,298 (deepspeed_trainer:226) INFO: 1epoch:train:11501-11600batch: iter_time=1.122e-04, loss_ctc=262.288, loss_att=229.977, acc=0.215, loss=239.705, grad_norm=11.476, loss_scale=1.000, learning_rate=1.925e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 16:59:43,270 (deepspeed_trainer:226) INFO: 1epoch:train:11601-11700batch: iter_time=1.103e-04, loss_ctc=267.173, loss_att=224.805, acc=0.202, loss=237.595, grad_norm=16.341, loss_scale=1.000, learning_rate=1.942e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:00:20,184 (deepspeed_trainer:226) INFO: 1epoch:train:11701-11800batch: iter_time=1.110e-04, loss_ctc=274.620, loss_att=231.273, acc=0.208, loss=244.360, grad_norm=10.032, loss_scale=1.000, learning_rate=1.959e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 17:00:58,287 (deepspeed_trainer:226) INFO: 1epoch:train:11801-11900batch: iter_time=1.084e-04, loss_ctc=297.467, loss_att=247.643, acc=0.199, loss=262.605, grad_norm=8.892, loss_scale=1.000, learning_rate=1.975e-05, step_time=0.380 [2024-11-28 17:01:35,581] [INFO] [logging.py:129:log_dist] [Rank 0] step=12000, skipped=0, lr=[np.float64(2.000166666666667e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:01:35,581] [INFO] [timer.py:264:stop] epoch=0/micro_step=12000/global_step=12000, RunningAvgSamplesPerSec=42.57873424977306, CurrSamplesPerSec=43.6582869324246, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 17:01:35,582 (deepspeed_trainer:226) INFO: 1epoch:train:11901-12000batch: iter_time=1.085e-04, loss_ctc=266.048, loss_att=229.162, acc=0.204, loss=240.188, grad_norm=12.318, loss_scale=1.000, learning_rate=1.992e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:02:13,349 (deepspeed_trainer:226) INFO: 1epoch:train:12001-12100batch: iter_time=1.144e-04, loss_ctc=272.895, loss_att=236.450, acc=0.205, loss=247.377, grad_norm=9.886, loss_scale=1.000, learning_rate=2.009e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 17:02:53,335 (deepspeed_trainer:226) INFO: 1epoch:train:12101-12200batch: iter_time=1.120e-04, loss_ctc=268.755, loss_att=225.278, acc=0.208, loss=238.345, grad_norm=9.207, loss_scale=1.000, learning_rate=2.025e-05, step_time=0.400 [cnode7-008:0/16] 2024-11-28 17:03:31,712 (deepspeed_trainer:226) INFO: 1epoch:train:12201-12300batch: iter_time=1.105e-04, loss_ctc=292.070, loss_att=238.213, acc=0.209, loss=254.320, grad_norm=10.627, loss_scale=1.000, learning_rate=2.042e-05, step_time=0.383 [cnode7-008:0/16] 2024-11-28 17:04:08,999 (deepspeed_trainer:226) INFO: 1epoch:train:12301-12400batch: iter_time=1.110e-04, loss_ctc=279.677, loss_att=246.293, acc=0.202, loss=256.322, grad_norm=11.645, loss_scale=1.000, learning_rate=2.059e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:04:46,254 (deepspeed_trainer:226) INFO: 1epoch:train:12401-12500batch: iter_time=1.140e-04, loss_ctc=268.910, loss_att=228.137, acc=0.207, loss=240.380, grad_norm=12.762, loss_scale=1.000, learning_rate=2.075e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:05:23,421 (deepspeed_trainer:226) INFO: 1epoch:train:12501-12600batch: iter_time=1.179e-04, loss_ctc=268.735, loss_att=224.347, acc=0.203, loss=237.675, grad_norm=14.082, loss_scale=1.000, learning_rate=2.092e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:06:00,745 (deepspeed_trainer:226) INFO: 1epoch:train:12601-12700batch: iter_time=1.101e-04, loss_ctc=278.329, loss_att=238.217, acc=0.207, loss=250.235, grad_norm=11.841, loss_scale=1.000, learning_rate=2.109e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:06:37,497 (deepspeed_trainer:226) INFO: 1epoch:train:12701-12800batch: iter_time=1.156e-04, loss_ctc=240.920, loss_att=198.205, acc=0.219, loss=211.032, grad_norm=10.986, loss_scale=1.000, learning_rate=2.125e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 17:07:14,790 (deepspeed_trainer:226) INFO: 1epoch:train:12801-12900batch: iter_time=1.159e-04, loss_ctc=286.728, loss_att=247.218, acc=0.206, loss=258.980, grad_norm=10.002, loss_scale=1.000, learning_rate=2.142e-05, step_time=0.373 [2024-11-28 17:07:52,006] [INFO] [logging.py:129:log_dist] [Rank 0] step=13000, skipped=0, lr=[np.float64(2.1668333333333334e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:07:52,007] [INFO] [timer.py:264:stop] epoch=0/micro_step=13000/global_step=13000, RunningAvgSamplesPerSec=42.63884479810395, CurrSamplesPerSec=41.97551345603025, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 17:07:52,009 (deepspeed_trainer:226) INFO: 1epoch:train:12901-13000batch: iter_time=1.115e-04, loss_ctc=284.498, loss_att=235.310, acc=0.208, loss=250.032, grad_norm=14.471, loss_scale=1.000, learning_rate=2.159e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:08:29,279 (deepspeed_trainer:226) INFO: 1epoch:train:13001-13100batch: iter_time=1.112e-04, loss_ctc=289.940, loss_att=239.062, acc=0.207, loss=254.330, grad_norm=9.786, loss_scale=1.000, learning_rate=2.175e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:08:42,376 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-28 17:09:08,936 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 17:09:15,106 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 17:09:15,106 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-28 17:09:15,108 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 17:09:42,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:42,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:42,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:43,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:40,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:40,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:43,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:44,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:44,169] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:44,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:09:41,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:30,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:30,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:32,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:29,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:32,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:32,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:29,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:30,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:10:31,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:16,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:18,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:19,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:19,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:19,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:19,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:20,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:20,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:17,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:17,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:19,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:20,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:20,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:20,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:21,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:11:23,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:03,550] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:05,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:06,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:06,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:07,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:07,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:08,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:08,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:05,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:06,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:07,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:08,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:08,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:09,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:09,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:12:13,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 17:13:13,892 (deepspeed_trainer:226) INFO: 1epoch:train:13101-13200batch: iter_time=2.364, loss_ctc=261.745, loss_att=227.762, acc=0.208, loss=237.895, grad_norm=14.104, loss_scale=1.000, learning_rate=2.192e-05, step_time=0.482 [cnode7-008:0/16] 2024-11-28 17:13:53,491 (deepspeed_trainer:226) INFO: 1epoch:train:13201-13300batch: iter_time=1.100e-04, loss_ctc=259.775, loss_att=223.482, acc=0.206, loss=234.222, grad_norm=9.475, loss_scale=1.000, learning_rate=2.209e-05, step_time=0.395 [cnode7-008:0/16] 2024-11-28 17:14:31,140 (deepspeed_trainer:226) INFO: 1epoch:train:13301-13400batch: iter_time=1.066e-04, loss_ctc=252.070, loss_att=216.425, acc=0.218, loss=227.050, grad_norm=12.815, loss_scale=1.000, learning_rate=2.225e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:15:08,326 (deepspeed_trainer:226) INFO: 1epoch:train:13401-13500batch: iter_time=1.098e-04, loss_ctc=255.553, loss_att=224.225, acc=0.215, loss=233.602, grad_norm=8.895, loss_scale=1.000, learning_rate=2.242e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:15:45,338 (deepspeed_trainer:226) INFO: 1epoch:train:13501-13600batch: iter_time=1.086e-04, loss_ctc=262.905, loss_att=222.143, acc=0.206, loss=234.475, grad_norm=16.503, loss_scale=1.000, learning_rate=2.259e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:16:22,538 (deepspeed_trainer:226) INFO: 1epoch:train:13601-13700batch: iter_time=1.077e-04, loss_ctc=282.455, loss_att=239.490, acc=0.214, loss=252.270, grad_norm=10.573, loss_scale=1.000, learning_rate=2.275e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:16:59,514 (deepspeed_trainer:226) INFO: 1epoch:train:13701-13800batch: iter_time=1.133e-04, loss_ctc=277.923, loss_att=233.887, acc=0.202, loss=247.192, grad_norm=9.812, loss_scale=1.000, learning_rate=2.292e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 17:17:36,965 (deepspeed_trainer:226) INFO: 1epoch:train:13801-13900batch: iter_time=1.095e-04, loss_ctc=290.690, loss_att=250.708, acc=0.206, loss=262.748, grad_norm=12.594, loss_scale=1.000, learning_rate=2.309e-05, step_time=0.374 [2024-11-28 17:18:14,088] [INFO] [logging.py:129:log_dist] [Rank 0] step=14000, skipped=0, lr=[np.float64(2.3335e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:18:14,089] [INFO] [timer.py:264:stop] epoch=0/micro_step=14000/global_step=14000, RunningAvgSamplesPerSec=42.61307604545249, CurrSamplesPerSec=46.06814924438594, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 17:18:14,090 (deepspeed_trainer:226) INFO: 1epoch:train:13901-14000batch: iter_time=1.089e-04, loss_ctc=256.892, loss_att=220.757, acc=0.209, loss=231.600, grad_norm=12.937, loss_scale=1.000, learning_rate=2.325e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 17:18:51,226 (deepspeed_trainer:226) INFO: 1epoch:train:14001-14100batch: iter_time=1.101e-04, loss_ctc=299.582, loss_att=245.940, acc=0.201, loss=262.043, grad_norm=13.937, loss_scale=1.000, learning_rate=2.342e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 17:19:28,233 (deepspeed_trainer:226) INFO: 1epoch:train:14101-14200batch: iter_time=1.096e-04, loss_ctc=256.212, loss_att=214.160, acc=0.219, loss=226.730, grad_norm=10.409, loss_scale=1.000, learning_rate=2.359e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:20:05,515 (deepspeed_trainer:226) INFO: 1epoch:train:14201-14300batch: iter_time=1.100e-04, loss_ctc=268.169, loss_att=233.956, acc=0.205, loss=244.185, grad_norm=7.972, loss_scale=1.000, learning_rate=2.375e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:20:43,274 (deepspeed_trainer:226) INFO: 1epoch:train:14301-14400batch: iter_time=1.085e-04, loss_ctc=283.955, loss_att=241.750, acc=0.206, loss=254.468, grad_norm=11.018, loss_scale=1.000, learning_rate=2.392e-05, step_time=0.378 [cnode7-008:0/16] 2024-11-28 17:21:20,824 (deepspeed_trainer:226) INFO: 1epoch:train:14401-14500batch: iter_time=1.044e-04, loss_ctc=264.603, loss_att=226.165, acc=0.207, loss=237.688, grad_norm=13.010, loss_scale=1.000, learning_rate=2.409e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 17:22:00,285 (deepspeed_trainer:226) INFO: 1epoch:train:14501-14600batch: iter_time=1.070e-04, loss_ctc=255.660, loss_att=223.223, acc=0.211, loss=232.926, grad_norm=12.084, loss_scale=1.000, learning_rate=2.425e-05, step_time=0.390 [cnode7-008:0/16] 2024-11-28 17:22:40,259 (deepspeed_trainer:226) INFO: 1epoch:train:14601-14700batch: iter_time=1.060e-04, loss_ctc=270.095, loss_att=227.377, acc=0.214, loss=240.123, grad_norm=9.299, loss_scale=1.000, learning_rate=2.442e-05, step_time=0.403 [cnode7-008:0/16] 2024-11-28 17:23:18,559 (deepspeed_trainer:226) INFO: 1epoch:train:14701-14800batch: iter_time=1.100e-04, loss_ctc=274.955, loss_att=235.287, acc=0.211, loss=247.210, grad_norm=10.403, loss_scale=1.000, learning_rate=2.459e-05, step_time=0.383 [cnode7-008:0/16] 2024-11-28 17:23:55,962 (deepspeed_trainer:226) INFO: 1epoch:train:14801-14900batch: iter_time=1.098e-04, loss_ctc=279.605, loss_att=238.498, acc=0.206, loss=250.855, grad_norm=11.285, loss_scale=1.000, learning_rate=2.475e-05, step_time=0.374 [2024-11-28 17:24:33,104] [INFO] [logging.py:129:log_dist] [Rank 0] step=15000, skipped=0, lr=[np.float64(2.5001666666666668e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:24:33,105] [INFO] [timer.py:264:stop] epoch=0/micro_step=15000/global_step=15000, RunningAvgSamplesPerSec=42.6409355353537, CurrSamplesPerSec=43.998017040267825, MemAllocated=1.77GB, MaxMemAllocated=15.26GB [cnode7-008:0/16] 2024-11-28 17:24:33,106 (deepspeed_trainer:226) INFO: 1epoch:train:14901-15000batch: iter_time=1.077e-04, loss_ctc=277.728, loss_att=233.482, acc=0.213, loss=246.800, grad_norm=10.802, loss_scale=1.000, learning_rate=2.492e-05, step_time=0.371 [2024-11-28 17:24:47,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:44,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:44,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:44,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:44,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:48,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:44,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:45,037] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:48,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:48,740] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:48,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:48,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:45,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:45,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:49,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:49,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:02,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:24:59,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:00,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:00,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:04,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:01,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:01,218] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:01,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:04,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:04,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:05,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:01,883] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:01,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:05,825] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:05,883] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:05,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:18,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:15,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:16,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:20,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:16,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:16,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:20,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:20,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:17,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:17,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:20,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:17,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:21,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:17,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:21,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:21,824] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:33,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:31,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:31,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:35,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:32,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:32,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:35,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:36,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:36,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:32,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:32,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:36,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:33,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:33,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:37,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:37,447] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:25:49,667] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 1 is about to be saved! [2024-11-28 17:25:49,697] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/mp_rank_00_model_states.pt [2024-11-28 17:25:49,697] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/mp_rank_00_model_states.pt... [2024-11-28 17:25:51,647] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/mp_rank_00_model_states.pt. [2024-11-28 17:25:51,803] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,804] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,318] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,319] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,319] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,805] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,319] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,319] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,320] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,807] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,320] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-28 17:25:48,321] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,810] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,810] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,811] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-28 17:25:51,811] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-28 17:25:52,463] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-28 17:25:52,469] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-28 17:25:52,469] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:52,538] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-28 17:25:52,538] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-28 17:25:52,538] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-28 17:25:49,178] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-28 17:25:49,178] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-28 17:25:49,178] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-28 17:25:49,178] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:49,178] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:53,708] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-28 17:25:53,708] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-28 17:25:53,708] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:50,560] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-28 17:25:50,560] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-28 17:25:50,560] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:54,111] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-28 17:25:54,111] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-28 17:25:54,111] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:50,632] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-28 17:25:50,632] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-28 17:25:50,632] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:54,119] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-28 17:25:54,119] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-28 17:25:54,119] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:50,633] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-28 17:25:50,633] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-28 17:25:50,633] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:50,690] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-28 17:25:50,690] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-28 17:25:50,690] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:54,199] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-28 17:25:54,199] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-28 17:25:54,199] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:54,491] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-28 17:25:54,491] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-28 17:25:54,491] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [2024-11-28 17:25:54,538] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-28 17:25:54,538] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_1/1/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-28 17:25:54,538] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 1 is ready now! [cnode7-008:0/16] 2024-11-28 17:25:54,547 (deepspeed_trainer:158) INFO: 1epoch results: [train] iter_time=0.124, loss_ctc=397.465, loss_att=274.261, acc=0.160, loss=311.217, grad_norm=23.132, loss_scale=1.000, learning_rate=1.250e-05, step_time=0.385, time=2 hours, 7 minutes and 13.9 seconds, total_count=15001, gpu_max_cached_mem_GB=27.027, [valid] loss_ctc=74.000, cer_ctc=1.000, loss_att=70.500, acc=0.228, cer=0.935, wer=1.000, loss=71.500, time=1 minute and 8.12 seconds, total_count=1, gpu_max_cached_mem_GB=27.027 [cnode7-008:0/16] 2024-11-28 17:25:57,220 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-28 17:26:23,423 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 17:26:28,930 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 17:26:28,930 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-28 17:26:28,932 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 17:26:48,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:45,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:49,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:46,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:50,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:50,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:47,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:47,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:47,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:47,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:47,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:51,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:51,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:48,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:52,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:26:52,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:31,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:35,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:35,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:36,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:37,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:37,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:34,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:34,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:34,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:34,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:38,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:35,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:35,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:39,330] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:36,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:27:40,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:21,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:18,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:23,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:23,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:23,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:24,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:24,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:25,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:21,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:22,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:22,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:22,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:22,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:23,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:27,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:28:25,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:05,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:09,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:09,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,836] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:11,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:12,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:13,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:10,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:11,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:11,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:11,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:29:15,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) /mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py:279: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with autocast(False): [cnode7-008:0/16] 2024-11-28 17:30:22,888 (deepspeed_trainer:226) INFO: 2epoch:train:1-100batch: iter_time=2.201, loss_ctc=262.415, loss_att=223.458, acc=0.219, loss=235.175, grad_norm=8.591, loss_scale=1.000, learning_rate=2.509e-05, step_time=0.457 [cnode7-008:0/16] 2024-11-28 17:31:00,194 (deepspeed_trainer:226) INFO: 2epoch:train:101-200batch: iter_time=1.035e-04, loss_ctc=273.430, loss_att=224.460, acc=0.212, loss=239.130, grad_norm=10.065, loss_scale=1.000, learning_rate=2.525e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:31:37,461 (deepspeed_trainer:226) INFO: 2epoch:train:201-300batch: iter_time=1.049e-04, loss_ctc=285.973, loss_att=241.765, acc=0.218, loss=255.093, grad_norm=9.241, loss_scale=1.000, learning_rate=2.542e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:32:15,096 (deepspeed_trainer:226) INFO: 2epoch:train:301-400batch: iter_time=1.077e-04, loss_ctc=268.476, loss_att=223.565, acc=0.219, loss=237.107, grad_norm=8.103, loss_scale=1.000, learning_rate=2.559e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:32:52,604 (deepspeed_trainer:226) INFO: 2epoch:train:401-500batch: iter_time=1.101e-04, loss_ctc=233.180, loss_att=195.730, acc=0.218, loss=206.993, grad_norm=8.123, loss_scale=1.000, learning_rate=2.575e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 17:33:30,258 (deepspeed_trainer:226) INFO: 2epoch:train:501-600batch: iter_time=1.060e-04, loss_ctc=265.444, loss_att=219.580, acc=0.227, loss=233.247, grad_norm=10.775, loss_scale=1.000, learning_rate=2.592e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:34:09,885 (deepspeed_trainer:226) INFO: 2epoch:train:601-700batch: iter_time=1.057e-04, loss_ctc=276.303, loss_att=220.398, acc=0.220, loss=237.202, grad_norm=10.950, loss_scale=1.000, learning_rate=2.609e-05, step_time=0.396 [cnode7-008:0/16] 2024-11-28 17:34:49,302 (deepspeed_trainer:226) INFO: 2epoch:train:701-800batch: iter_time=1.056e-04, loss_ctc=240.368, loss_att=199.042, acc=0.225, loss=211.468, grad_norm=10.101, loss_scale=1.000, learning_rate=2.625e-05, step_time=0.394 [cnode7-008:0/16] 2024-11-28 17:35:27,161 (deepspeed_trainer:226) INFO: 2epoch:train:801-900batch: iter_time=1.091e-04, loss_ctc=279.277, loss_att=226.755, acc=0.228, loss=242.510, grad_norm=9.443, loss_scale=1.000, learning_rate=2.642e-05, step_time=0.378 [2024-11-28 17:36:04,678] [INFO] [logging.py:129:log_dist] [Rank 0] step=16000, skipped=0, lr=[np.float64(2.6668333333333334e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:36:04,678] [INFO] [timer.py:264:stop] epoch=0/micro_step=16000/global_step=16000, RunningAvgSamplesPerSec=42.60506107205661, CurrSamplesPerSec=43.90349014125523, MemAllocated=2.04GB, MaxMemAllocated=15.52GB [cnode7-008:0/16] 2024-11-28 17:36:04,680 (deepspeed_trainer:226) INFO: 2epoch:train:901-1000batch: iter_time=1.066e-04, loss_ctc=262.735, loss_att=211.895, acc=0.220, loss=227.240, grad_norm=11.006, loss_scale=1.000, learning_rate=2.659e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 17:36:42,006 (deepspeed_trainer:226) INFO: 2epoch:train:1001-1100batch: iter_time=1.057e-04, loss_ctc=249.450, loss_att=210.643, acc=0.229, loss=222.238, grad_norm=10.746, loss_scale=1.000, learning_rate=2.675e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:37:19,700 (deepspeed_trainer:226) INFO: 2epoch:train:1101-1200batch: iter_time=1.056e-04, loss_ctc=285.290, loss_att=224.423, acc=0.206, loss=242.702, grad_norm=10.989, loss_scale=1.000, learning_rate=2.692e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 17:37:57,391 (deepspeed_trainer:226) INFO: 2epoch:train:1201-1300batch: iter_time=1.058e-04, loss_ctc=244.798, loss_att=197.015, acc=0.225, loss=211.347, grad_norm=8.587, loss_scale=1.000, learning_rate=2.709e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 17:38:35,000 (deepspeed_trainer:226) INFO: 2epoch:train:1301-1400batch: iter_time=1.030e-04, loss_ctc=274.522, loss_att=225.643, acc=0.226, loss=240.260, grad_norm=8.666, loss_scale=1.000, learning_rate=2.725e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:39:13,158 (deepspeed_trainer:226) INFO: 2epoch:train:1401-1500batch: iter_time=1.060e-04, loss_ctc=267.017, loss_att=216.438, acc=0.219, loss=231.673, grad_norm=7.750, loss_scale=1.000, learning_rate=2.742e-05, step_time=0.381 [cnode7-008:0/16] 2024-11-28 17:39:50,699 (deepspeed_trainer:226) INFO: 2epoch:train:1501-1600batch: iter_time=1.063e-04, loss_ctc=255.558, loss_att=212.398, acc=0.222, loss=225.333, grad_norm=8.425, loss_scale=1.000, learning_rate=2.759e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 17:40:28,379 (deepspeed_trainer:226) INFO: 2epoch:train:1601-1700batch: iter_time=1.096e-04, loss_ctc=251.773, loss_att=211.137, acc=0.222, loss=223.322, grad_norm=10.600, loss_scale=1.000, learning_rate=2.775e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:41:06,969 (deepspeed_trainer:226) INFO: 2epoch:train:1701-1800batch: iter_time=1.117e-04, loss_ctc=268.675, loss_att=219.637, acc=0.223, loss=234.373, grad_norm=8.613, loss_scale=1.000, learning_rate=2.792e-05, step_time=0.386 [cnode7-008:0/16] 2024-11-28 17:41:41,058 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-28 17:42:06,687 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 17:42:12,102 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 17:42:12,102 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-28 17:42:12,217 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 17:42:44,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:44,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:44,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:44,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:44,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:45,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:45,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:45,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:41,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:42:42,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:36,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:33,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:37,461] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:37,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:34,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:38,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:34,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:38,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:34,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:38,643] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:38,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:38,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:35,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:35,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:35,377] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:43:37,067] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:23,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:21,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:25,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:25,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:22,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:25,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:22,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:25,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:26,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:22,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:23,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:23,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:28,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:26,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:30,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:44:27,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:10,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:11,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:13,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:09,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:13,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:13,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:10,120] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:13,877] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:10,450] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:10,698] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:11,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:16,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:15,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:19,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:16,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:45:17,296] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 17:45:58,745 (deepspeed_trainer:226) INFO: 2epoch:train:1801-1900batch: iter_time=2.420, loss_ctc=271.717, loss_att=217.903, acc=0.225, loss=234.010, grad_norm=8.497, loss_scale=1.000, learning_rate=2.809e-05, step_time=0.498 [2024-11-28 17:46:35,697] [INFO] [logging.py:129:log_dist] [Rank 0] step=17000, skipped=0, lr=[np.float64(2.8335e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:46:35,700] [INFO] [timer.py:264:stop] epoch=0/micro_step=17000/global_step=17000, RunningAvgSamplesPerSec=42.562414205321005, CurrSamplesPerSec=44.065244357764385, MemAllocated=2.04GB, MaxMemAllocated=15.52GB [cnode7-008:0/16] 2024-11-28 17:46:35,702 (deepspeed_trainer:226) INFO: 2epoch:train:1901-2000batch: iter_time=1.023e-04, loss_ctc=248.697, loss_att=208.317, acc=0.225, loss=220.363, grad_norm=8.678, loss_scale=1.000, learning_rate=2.825e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 17:47:12,770 (deepspeed_trainer:226) INFO: 2epoch:train:2001-2100batch: iter_time=1.042e-04, loss_ctc=281.655, loss_att=230.710, acc=0.211, loss=245.972, grad_norm=10.243, loss_scale=1.000, learning_rate=2.842e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:47:49,562 (deepspeed_trainer:226) INFO: 2epoch:train:2101-2200batch: iter_time=1.077e-04, loss_ctc=252.669, loss_att=206.481, acc=0.232, loss=220.290, grad_norm=8.210, loss_scale=1.000, learning_rate=2.859e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 17:48:27,154 (deepspeed_trainer:226) INFO: 2epoch:train:2201-2300batch: iter_time=1.078e-04, loss_ctc=269.955, loss_att=222.643, acc=0.222, loss=236.830, grad_norm=9.559, loss_scale=1.000, learning_rate=2.875e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:49:04,210 (deepspeed_trainer:226) INFO: 2epoch:train:2301-2400batch: iter_time=1.058e-04, loss_ctc=245.778, loss_att=198.876, acc=0.227, loss=212.993, grad_norm=8.567, loss_scale=1.000, learning_rate=2.892e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:49:41,590 (deepspeed_trainer:226) INFO: 2epoch:train:2401-2500batch: iter_time=1.076e-04, loss_ctc=266.415, loss_att=219.683, acc=0.226, loss=233.648, grad_norm=8.852, loss_scale=1.000, learning_rate=2.909e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:50:19,111 (deepspeed_trainer:226) INFO: 2epoch:train:2501-2600batch: iter_time=1.140e-04, loss_ctc=277.550, loss_att=221.105, acc=0.229, loss=238.015, grad_norm=10.861, loss_scale=1.000, learning_rate=2.925e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 17:50:56,599 (deepspeed_trainer:226) INFO: 2epoch:train:2601-2700batch: iter_time=1.091e-04, loss_ctc=230.935, loss_att=185.648, acc=0.234, loss=199.282, grad_norm=9.416, loss_scale=1.000, learning_rate=2.942e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 17:51:34,790 (deepspeed_trainer:226) INFO: 2epoch:train:2701-2800batch: iter_time=1.087e-04, loss_ctc=284.353, loss_att=225.305, acc=0.226, loss=243.002, grad_norm=8.301, loss_scale=1.000, learning_rate=2.959e-05, step_time=0.381 [cnode7-008:0/16] 2024-11-28 17:52:13,614 (deepspeed_trainer:226) INFO: 2epoch:train:2801-2900batch: iter_time=1.085e-04, loss_ctc=243.065, loss_att=201.195, acc=0.229, loss=213.885, grad_norm=8.383, loss_scale=1.000, learning_rate=2.975e-05, step_time=0.388 [2024-11-28 17:52:53,058] [INFO] [logging.py:129:log_dist] [Rank 0] step=18000, skipped=0, lr=[np.float64(3.0001666666666668e-05)], mom=[[0.9, 0.98]] [2024-11-28 17:52:53,059] [INFO] [timer.py:264:stop] epoch=0/micro_step=18000/global_step=18000, RunningAvgSamplesPerSec=42.59962293284277, CurrSamplesPerSec=43.528223804326565, MemAllocated=2.04GB, MaxMemAllocated=15.52GB [cnode7-008:0/16] 2024-11-28 17:52:53,060 (deepspeed_trainer:226) INFO: 2epoch:train:2901-3000batch: iter_time=1.093e-04, loss_ctc=242.225, loss_att=198.960, acc=0.234, loss=211.952, grad_norm=11.697, loss_scale=1.000, learning_rate=2.992e-05, step_time=0.394 [cnode7-008:0/16] 2024-11-28 17:53:30,232 (deepspeed_trainer:226) INFO: 2epoch:train:3001-3100batch: iter_time=1.101e-04, loss_ctc=299.137, loss_att=229.977, acc=0.213, loss=250.798, grad_norm=10.565, loss_scale=1.000, learning_rate=3.009e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 17:54:07,029 (deepspeed_trainer:226) INFO: 2epoch:train:3101-3200batch: iter_time=1.104e-04, loss_ctc=250.580, loss_att=200.840, acc=0.223, loss=215.700, grad_norm=6.625, loss_scale=1.000, learning_rate=3.025e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 17:54:44,660 (deepspeed_trainer:226) INFO: 2epoch:train:3201-3300batch: iter_time=1.117e-04, loss_ctc=270.725, loss_att=220.350, acc=0.232, loss=235.417, grad_norm=8.505, loss_scale=1.000, learning_rate=3.042e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 17:55:21,931 (deepspeed_trainer:226) INFO: 2epoch:train:3301-3400batch: iter_time=1.166e-04, loss_ctc=261.105, loss_att=210.625, acc=0.226, loss=225.863, grad_norm=9.382, loss_scale=1.000, learning_rate=3.059e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 17:55:59,075 (deepspeed_trainer:226) INFO: 2epoch:train:3401-3500batch: iter_time=1.146e-04, loss_ctc=246.463, loss_att=201.905, acc=0.224, loss=215.355, grad_norm=8.777, loss_scale=1.000, learning_rate=3.075e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 17:56:37,110 (deepspeed_trainer:226) INFO: 2epoch:train:3501-3600batch: iter_time=1.147e-04, loss_ctc=251.855, loss_att=208.032, acc=0.226, loss=221.123, grad_norm=8.853, loss_scale=1.000, learning_rate=3.092e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 17:57:14,157 (deepspeed_trainer:226) INFO: 2epoch:train:3601-3700batch: iter_time=1.124e-04, loss_ctc=269.892, loss_att=216.613, acc=0.224, loss=232.640, grad_norm=8.573, loss_scale=1.000, learning_rate=3.109e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 17:57:35,831 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-28 17:58:02,533 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 17:58:08,011 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 17:58:08,012 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-28 17:58:08,014 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 17:58:34,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:34,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:31,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:35,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:31,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:35,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:35,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:36,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:36,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,548] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:36,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:58:32,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:22,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:22,280] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:19,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:23,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:23,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:19,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:19,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:23,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:23,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:20,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:20,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:20,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:21,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:21,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:26,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 17:59:27,157] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:08,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:08,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:10,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:10,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:06,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:11,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:11,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:08,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:08,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:08,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:09,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:09,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:14,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:16,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:12,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:13,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:55,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:55,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:53,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:57,440] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:57,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:58,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:59,197] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:56,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:56,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:56,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:56,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:00:57,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:01:02,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:01:05,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:01:02,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:01:02,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 18:01:52,313 (deepspeed_trainer:226) INFO: 2epoch:train:3701-3800batch: iter_time=2.357, loss_ctc=266.762, loss_att=218.225, acc=0.231, loss=232.852, grad_norm=8.089, loss_scale=1.000, learning_rate=3.125e-05, step_time=0.424 [cnode7-008:0/16] 2024-11-28 18:02:30,683 (deepspeed_trainer:226) INFO: 2epoch:train:3801-3900batch: iter_time=1.061e-04, loss_ctc=254.012, loss_att=203.083, acc=0.227, loss=218.310, grad_norm=8.605, loss_scale=1.000, learning_rate=3.142e-05, step_time=0.383 [2024-11-28 18:03:08,177] [INFO] [logging.py:129:log_dist] [Rank 0] step=19000, skipped=0, lr=[np.float64(3.1668333333333336e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:03:08,178] [INFO] [timer.py:264:stop] epoch=0/micro_step=19000/global_step=19000, RunningAvgSamplesPerSec=42.61980015124136, CurrSamplesPerSec=44.99900107119304, MemAllocated=2.04GB, MaxMemAllocated=15.52GB [cnode7-008:0/16] 2024-11-28 18:03:08,179 (deepspeed_trainer:226) INFO: 2epoch:train:3901-4000batch: iter_time=1.060e-04, loss_ctc=274.080, loss_att=223.877, acc=0.217, loss=238.960, grad_norm=7.588, loss_scale=1.000, learning_rate=3.159e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 18:03:45,508 (deepspeed_trainer:226) INFO: 2epoch:train:4001-4100batch: iter_time=1.082e-04, loss_ctc=263.430, loss_att=215.532, acc=0.234, loss=229.912, grad_norm=9.046, loss_scale=1.000, learning_rate=3.175e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:04:22,294 (deepspeed_trainer:226) INFO: 2epoch:train:4101-4200batch: iter_time=1.115e-04, loss_ctc=242.763, loss_att=195.551, acc=0.227, loss=209.591, grad_norm=7.720, loss_scale=1.000, learning_rate=3.192e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 18:04:59,278 (deepspeed_trainer:226) INFO: 2epoch:train:4201-4300batch: iter_time=1.074e-04, loss_ctc=261.731, loss_att=214.644, acc=0.236, loss=228.799, grad_norm=7.689, loss_scale=1.000, learning_rate=3.209e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 18:05:36,272 (deepspeed_trainer:226) INFO: 2epoch:train:4301-4400batch: iter_time=1.101e-04, loss_ctc=269.128, loss_att=209.826, acc=0.226, loss=227.741, grad_norm=9.548, loss_scale=1.000, learning_rate=3.225e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 18:06:13,189 (deepspeed_trainer:226) INFO: 2epoch:train:4401-4500batch: iter_time=1.085e-04, loss_ctc=253.778, loss_att=204.055, acc=0.236, loss=219.023, grad_norm=7.211, loss_scale=1.000, learning_rate=3.242e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 18:06:50,525 (deepspeed_trainer:226) INFO: 2epoch:train:4501-4600batch: iter_time=1.042e-04, loss_ctc=233.060, loss_att=188.417, acc=0.242, loss=201.745, grad_norm=7.053, loss_scale=1.000, learning_rate=3.259e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:07:27,851 (deepspeed_trainer:226) INFO: 2epoch:train:4601-4700batch: iter_time=1.061e-04, loss_ctc=293.277, loss_att=230.722, acc=0.225, loss=249.423, grad_norm=8.291, loss_scale=1.000, learning_rate=3.275e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:08:04,691 (deepspeed_trainer:226) INFO: 2epoch:train:4701-4800batch: iter_time=1.091e-04, loss_ctc=232.893, loss_att=187.322, acc=0.239, loss=201.060, grad_norm=7.366, loss_scale=1.000, learning_rate=3.292e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 18:08:42,300 (deepspeed_trainer:226) INFO: 2epoch:train:4801-4900batch: iter_time=1.138e-04, loss_ctc=274.478, loss_att=218.315, acc=0.220, loss=235.118, grad_norm=9.072, loss_scale=1.000, learning_rate=3.309e-05, step_time=0.376 [2024-11-28 18:09:19,747] [INFO] [logging.py:129:log_dist] [Rank 0] step=20000, skipped=0, lr=[np.float64(3.3335e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:09:19,747] [INFO] [timer.py:264:stop] epoch=0/micro_step=20000/global_step=20000, RunningAvgSamplesPerSec=42.682973498573, CurrSamplesPerSec=43.41033104916819, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:09:19,749 (deepspeed_trainer:226) INFO: 2epoch:train:4901-5000batch: iter_time=1.193e-04, loss_ctc=270.291, loss_att=208.908, acc=0.230, loss=227.333, grad_norm=8.003, loss_scale=1.000, learning_rate=3.325e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 18:09:59,437 (deepspeed_trainer:226) INFO: 2epoch:train:5001-5100batch: iter_time=1.108e-04, loss_ctc=243.512, loss_att=195.602, acc=0.235, loss=210.030, grad_norm=8.176, loss_scale=1.000, learning_rate=3.342e-05, step_time=0.397 [cnode7-008:0/16] 2024-11-28 18:10:38,478 (deepspeed_trainer:226) INFO: 2epoch:train:5101-5200batch: iter_time=1.151e-04, loss_ctc=283.205, loss_att=226.333, acc=0.228, loss=243.390, grad_norm=8.828, loss_scale=1.000, learning_rate=3.359e-05, step_time=0.390 [cnode7-008:0/16] 2024-11-28 18:11:16,533 (deepspeed_trainer:226) INFO: 2epoch:train:5201-5300batch: iter_time=1.151e-04, loss_ctc=237.882, loss_att=188.560, acc=0.234, loss=203.338, grad_norm=7.553, loss_scale=1.000, learning_rate=3.375e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 18:11:53,867 (deepspeed_trainer:226) INFO: 2epoch:train:5301-5400batch: iter_time=1.160e-04, loss_ctc=247.732, loss_att=202.555, acc=0.233, loss=216.148, grad_norm=8.030, loss_scale=1.000, learning_rate=3.392e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:12:30,994 (deepspeed_trainer:226) INFO: 2epoch:train:5401-5500batch: iter_time=1.165e-04, loss_ctc=264.738, loss_att=211.255, acc=0.226, loss=227.400, grad_norm=7.742, loss_scale=1.000, learning_rate=3.409e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 18:13:08,240 (deepspeed_trainer:226) INFO: 2epoch:train:5501-5600batch: iter_time=1.165e-04, loss_ctc=261.752, loss_att=209.488, acc=0.232, loss=225.265, grad_norm=7.725, loss_scale=1.000, learning_rate=3.425e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:13:20,535 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-28 18:13:46,995 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 18:13:52,632 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 18:13:52,632 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-28 18:13:52,635 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 18:14:19,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:19,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:16,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:16,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:16,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:20,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:20,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:17,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:21,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:21,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:21,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:21,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:17,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:17,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:17,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:14:18,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:08,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:04,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:08,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:08,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:04,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:04,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:09,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:09,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:05,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:09,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:09,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:09,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:06,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:06,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:06,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:06,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:56,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:56,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:57,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:53,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:53,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:57,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:53,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:58,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:54,005] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:58,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:58,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:58,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:54,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:54,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:55,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:15:57,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:44,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:44,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:45,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:45,442] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:45,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:41,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:42,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:42,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:42,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:43,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:43,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:43,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:48,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:49,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:50,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:16:47,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 18:17:46,069 (deepspeed_trainer:226) INFO: 2epoch:train:5601-5700batch: iter_time=2.393, loss_ctc=261.655, loss_att=220.945, acc=0.234, loss=233.165, grad_norm=7.782, loss_scale=1.000, learning_rate=3.442e-05, step_time=0.384 [cnode7-008:0/16] 2024-11-28 18:18:23,951 (deepspeed_trainer:226) INFO: 2epoch:train:5701-5800batch: iter_time=1.135e-04, loss_ctc=257.738, loss_att=213.735, acc=0.220, loss=226.985, grad_norm=9.953, loss_scale=1.000, learning_rate=3.459e-05, step_time=0.379 [cnode7-008:0/16] 2024-11-28 18:19:02,062 (deepspeed_trainer:226) INFO: 2epoch:train:5801-5900batch: iter_time=1.145e-04, loss_ctc=278.805, loss_att=236.105, acc=0.222, loss=248.970, grad_norm=8.041, loss_scale=1.000, learning_rate=3.475e-05, step_time=0.380 [2024-11-28 18:19:40,911] [INFO] [logging.py:129:log_dist] [Rank 0] step=21000, skipped=0, lr=[np.float64(3.500166666666667e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:19:40,912] [INFO] [timer.py:264:stop] epoch=0/micro_step=21000/global_step=21000, RunningAvgSamplesPerSec=42.68442295717337, CurrSamplesPerSec=43.6891820791335, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:19:40,913 (deepspeed_trainer:226) INFO: 2epoch:train:5901-6000batch: iter_time=1.085e-04, loss_ctc=240.995, loss_att=197.001, acc=0.235, loss=210.233, grad_norm=7.346, loss_scale=1.000, learning_rate=3.492e-05, step_time=0.388 [cnode7-008:0/16] 2024-11-28 18:20:21,067 (deepspeed_trainer:226) INFO: 2epoch:train:6001-6100batch: iter_time=1.086e-04, loss_ctc=238.675, loss_att=191.145, acc=0.229, loss=205.495, grad_norm=7.887, loss_scale=1.000, learning_rate=3.509e-05, step_time=0.401 [cnode7-008:0/16] 2024-11-28 18:20:59,463 (deepspeed_trainer:226) INFO: 2epoch:train:6101-6200batch: iter_time=1.089e-04, loss_ctc=277.038, loss_att=222.630, acc=0.230, loss=238.879, grad_norm=7.195, loss_scale=1.000, learning_rate=3.525e-05, step_time=0.384 [cnode7-008:0/16] 2024-11-28 18:21:36,436 (deepspeed_trainer:226) INFO: 2epoch:train:6201-6300batch: iter_time=1.091e-04, loss_ctc=251.700, loss_att=204.718, acc=0.231, loss=218.800, grad_norm=8.014, loss_scale=1.000, learning_rate=3.542e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 18:22:13,387 (deepspeed_trainer:226) INFO: 2epoch:train:6301-6400batch: iter_time=1.085e-04, loss_ctc=248.343, loss_att=202.665, acc=0.238, loss=216.395, grad_norm=8.717, loss_scale=1.000, learning_rate=3.559e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 18:22:50,647 (deepspeed_trainer:226) INFO: 2epoch:train:6401-6500batch: iter_time=1.091e-04, loss_ctc=250.030, loss_att=202.260, acc=0.239, loss=216.590, grad_norm=6.850, loss_scale=1.000, learning_rate=3.575e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:23:27,951 (deepspeed_trainer:226) INFO: 2epoch:train:6501-6600batch: iter_time=1.191e-04, loss_ctc=278.048, loss_att=223.805, acc=0.226, loss=240.077, grad_norm=7.612, loss_scale=1.000, learning_rate=3.592e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:24:04,975 (deepspeed_trainer:226) INFO: 2epoch:train:6601-6700batch: iter_time=1.093e-04, loss_ctc=226.683, loss_att=188.385, acc=0.244, loss=199.895, grad_norm=6.997, loss_scale=1.000, learning_rate=3.609e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 18:24:42,352 (deepspeed_trainer:226) INFO: 2epoch:train:6701-6800batch: iter_time=1.095e-04, loss_ctc=277.420, loss_att=217.655, acc=0.218, loss=235.643, grad_norm=8.976, loss_scale=1.000, learning_rate=3.625e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:25:19,720 (deepspeed_trainer:226) INFO: 2epoch:train:6801-6900batch: iter_time=1.110e-04, loss_ctc=267.580, loss_att=214.952, acc=0.234, loss=230.673, grad_norm=7.846, loss_scale=1.000, learning_rate=3.642e-05, step_time=0.374 [2024-11-28 18:25:56,784] [INFO] [logging.py:129:log_dist] [Rank 0] step=22000, skipped=0, lr=[np.float64(3.6668333333333336e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:25:56,785] [INFO] [timer.py:264:stop] epoch=0/micro_step=22000/global_step=22000, RunningAvgSamplesPerSec=42.71712223100089, CurrSamplesPerSec=42.703993284358525, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:25:56,786 (deepspeed_trainer:226) INFO: 2epoch:train:6901-7000batch: iter_time=1.159e-04, loss_ctc=244.855, loss_att=198.045, acc=0.238, loss=212.115, grad_norm=7.378, loss_scale=1.000, learning_rate=3.659e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 18:26:34,802 (deepspeed_trainer:226) INFO: 2epoch:train:7001-7100batch: iter_time=1.081e-04, loss_ctc=270.507, loss_att=219.975, acc=0.236, loss=235.113, grad_norm=7.233, loss_scale=1.000, learning_rate=3.675e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 18:27:12,830 (deepspeed_trainer:226) INFO: 2epoch:train:7101-7200batch: iter_time=1.095e-04, loss_ctc=236.830, loss_att=189.925, acc=0.235, loss=203.998, grad_norm=10.188, loss_scale=1.000, learning_rate=3.692e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 18:27:50,699 (deepspeed_trainer:226) INFO: 2epoch:train:7201-7300batch: iter_time=1.100e-04, loss_ctc=247.667, loss_att=204.780, acc=0.233, loss=217.600, grad_norm=7.418, loss_scale=1.000, learning_rate=3.709e-05, step_time=0.378 [cnode7-008:0/16] 2024-11-28 18:28:31,428 (deepspeed_trainer:226) INFO: 2epoch:train:7301-7400batch: iter_time=1.094e-04, loss_ctc=281.558, loss_att=227.093, acc=0.225, loss=243.468, grad_norm=8.609, loss_scale=1.000, learning_rate=3.725e-05, step_time=0.407 [cnode7-008:0/16] 2024-11-28 18:29:11,314 (deepspeed_trainer:226) INFO: 2epoch:train:7401-7500batch: iter_time=1.104e-04, loss_ctc=240.708, loss_att=194.715, acc=0.238, loss=208.398, grad_norm=7.036, loss_scale=1.000, learning_rate=3.742e-05, step_time=0.399 [cnode7-008:0/16] 2024-11-28 18:29:14,744 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-28 18:29:41,407 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 18:29:47,004 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 18:29:47,004 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-28 18:29:47,006 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 18:30:13,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:10,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:10,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:14,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:10,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:10,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:15,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:15,610] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:15,815] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:15,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:11,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:11,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:15,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:11,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:16,066] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:11,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:57,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:01,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:02,159] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:58,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:02,604] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:58,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:02,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:03,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:03,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:59,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:30:59,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:04,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:00,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:00,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:00,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:07,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:43,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:48,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:49,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:49,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:45,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:50,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:46,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:50,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:50,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:47,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:47,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:51,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:48,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:48,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:52,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:31:57,170] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:31,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,973] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:37,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:34,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:38,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:34,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:39,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:35,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:36,284] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:37,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:45,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:32:42,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 18:33:49,271 (deepspeed_trainer:226) INFO: 2epoch:train:7501-7600batch: iter_time=2.314, loss_ctc=252.635, loss_att=212.600, acc=0.236, loss=224.588, grad_norm=6.113, loss_scale=1.000, learning_rate=3.759e-05, step_time=0.465 [cnode7-008:0/16] 2024-11-28 18:34:26,542 (deepspeed_trainer:226) INFO: 2epoch:train:7601-7700batch: iter_time=1.023e-04, loss_ctc=260.930, loss_att=214.655, acc=0.226, loss=228.577, grad_norm=8.041, loss_scale=1.000, learning_rate=3.775e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:35:04,029 (deepspeed_trainer:226) INFO: 2epoch:train:7701-7800batch: iter_time=1.066e-04, loss_ctc=276.658, loss_att=228.512, acc=0.235, loss=242.907, grad_norm=8.092, loss_scale=1.000, learning_rate=3.792e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 18:35:41,369 (deepspeed_trainer:226) INFO: 2epoch:train:7801-7900batch: iter_time=1.076e-04, loss_ctc=259.399, loss_att=208.668, acc=0.233, loss=223.917, grad_norm=6.471, loss_scale=1.000, learning_rate=3.809e-05, step_time=0.373 [2024-11-28 18:36:18,278] [INFO] [logging.py:129:log_dist] [Rank 0] step=23000, skipped=0, lr=[np.float64(3.8335e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:36:18,278] [INFO] [timer.py:264:stop] epoch=0/micro_step=23000/global_step=23000, RunningAvgSamplesPerSec=42.67629233695296, CurrSamplesPerSec=43.822699830657825, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:36:18,280 (deepspeed_trainer:226) INFO: 2epoch:train:7901-8000batch: iter_time=1.062e-04, loss_ctc=223.675, loss_att=181.463, acc=0.237, loss=194.062, grad_norm=6.988, loss_scale=1.000, learning_rate=3.825e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 18:36:56,316 (deepspeed_trainer:226) INFO: 2epoch:train:8001-8100batch: iter_time=1.063e-04, loss_ctc=258.216, loss_att=206.013, acc=0.243, loss=221.590, grad_norm=6.638, loss_scale=1.000, learning_rate=3.842e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 18:37:33,811 (deepspeed_trainer:226) INFO: 2epoch:train:8101-8200batch: iter_time=1.067e-04, loss_ctc=264.985, loss_att=213.270, acc=0.235, loss=228.782, grad_norm=7.715, loss_scale=1.000, learning_rate=3.859e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 18:38:11,537 (deepspeed_trainer:226) INFO: 2epoch:train:8201-8300batch: iter_time=1.067e-04, loss_ctc=229.865, loss_att=189.575, acc=0.244, loss=201.735, grad_norm=6.903, loss_scale=1.000, learning_rate=3.875e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 18:38:52,593 (deepspeed_trainer:226) INFO: 2epoch:train:8301-8400batch: iter_time=1.093e-04, loss_ctc=266.680, loss_att=214.930, acc=0.240, loss=230.415, grad_norm=6.535, loss_scale=1.000, learning_rate=3.892e-05, step_time=0.410 [cnode7-008:0/16] 2024-11-28 18:39:30,626 (deepspeed_trainer:226) INFO: 2epoch:train:8401-8500batch: iter_time=1.084e-04, loss_ctc=254.170, loss_att=200.925, acc=0.233, loss=216.840, grad_norm=6.564, loss_scale=1.000, learning_rate=3.909e-05, step_time=0.381 [cnode7-008:0/16] 2024-11-28 18:40:08,108 (deepspeed_trainer:226) INFO: 2epoch:train:8501-8600batch: iter_time=1.074e-04, loss_ctc=238.350, loss_att=200.315, acc=0.247, loss=211.605, grad_norm=6.907, loss_scale=1.000, learning_rate=3.925e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 18:40:45,795 (deepspeed_trainer:226) INFO: 2epoch:train:8601-8700batch: iter_time=1.070e-04, loss_ctc=272.832, loss_att=209.785, acc=0.228, loss=228.745, grad_norm=7.100, loss_scale=1.000, learning_rate=3.942e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 18:41:23,075 (deepspeed_trainer:226) INFO: 2epoch:train:8701-8800batch: iter_time=1.082e-04, loss_ctc=236.540, loss_att=186.915, acc=0.244, loss=201.873, grad_norm=6.662, loss_scale=1.000, learning_rate=3.959e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:42:00,473 (deepspeed_trainer:226) INFO: 2epoch:train:8801-8900batch: iter_time=1.077e-04, loss_ctc=262.315, loss_att=215.387, acc=0.241, loss=229.442, grad_norm=7.232, loss_scale=1.000, learning_rate=3.975e-05, step_time=0.374 [2024-11-28 18:42:37,691] [INFO] [logging.py:129:log_dist] [Rank 0] step=24000, skipped=0, lr=[np.float64(4.000166666666667e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:42:37,691] [INFO] [timer.py:264:stop] epoch=0/micro_step=24000/global_step=24000, RunningAvgSamplesPerSec=42.689457513229996, CurrSamplesPerSec=44.305047462436164, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:42:37,693 (deepspeed_trainer:226) INFO: 2epoch:train:8901-9000batch: iter_time=1.073e-04, loss_ctc=258.137, loss_att=202.820, acc=0.237, loss=219.440, grad_norm=7.071, loss_scale=1.000, learning_rate=3.992e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:43:14,953 (deepspeed_trainer:226) INFO: 2epoch:train:9001-9100batch: iter_time=1.086e-04, loss_ctc=247.445, loss_att=202.188, acc=0.238, loss=215.860, grad_norm=7.223, loss_scale=1.000, learning_rate=4.009e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:43:51,873 (deepspeed_trainer:226) INFO: 2epoch:train:9101-9200batch: iter_time=1.084e-04, loss_ctc=240.042, loss_att=197.252, acc=0.242, loss=210.055, grad_norm=6.768, loss_scale=1.000, learning_rate=4.025e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 18:44:29,341 (deepspeed_trainer:226) INFO: 2epoch:train:9201-9300batch: iter_time=1.080e-04, loss_ctc=260.955, loss_att=212.257, acc=0.238, loss=226.938, grad_norm=6.822, loss_scale=1.000, learning_rate=4.042e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 18:45:00,824 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-28 18:45:27,715 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 18:45:34,130 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 18:45:34,130 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-28 18:45:34,132 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 18:46:01,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:01,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:57,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:01,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:02,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:02,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:58,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:58,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:02,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:02,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:02,833] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:58,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:58,811] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:59,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:59,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:45:59,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:48,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:49,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:49,447] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:49,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:50,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:50,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:50,578] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:46,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:46,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:50,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:46,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:47,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:47,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:47,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:48,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:46:51,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:35,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:36,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:38,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:38,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:38,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:35,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:36,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:36,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:36,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,068] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:37,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:47:51,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:22,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:23,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:24,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:24,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:26,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:26,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:23,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:24,197] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:24,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:25,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 18:48:43,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 18:49:22,579 (deepspeed_trainer:226) INFO: 2epoch:train:9301-9400batch: iter_time=2.335, loss_ctc=257.715, loss_att=205.613, acc=0.240, loss=221.282, grad_norm=6.874, loss_scale=1.000, learning_rate=4.059e-05, step_time=0.597 [cnode7-008:0/16] 2024-11-28 18:50:00,943 (deepspeed_trainer:226) INFO: 2epoch:train:9401-9500batch: iter_time=1.009e-04, loss_ctc=239.195, loss_att=195.602, acc=0.241, loss=208.778, grad_norm=6.931, loss_scale=1.000, learning_rate=4.075e-05, step_time=0.383 [cnode7-008:0/16] 2024-11-28 18:50:38,359 (deepspeed_trainer:226) INFO: 2epoch:train:9501-9600batch: iter_time=1.016e-04, loss_ctc=265.457, loss_att=215.493, acc=0.226, loss=230.450, grad_norm=6.458, loss_scale=1.000, learning_rate=4.092e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 18:51:15,572 (deepspeed_trainer:226) INFO: 2epoch:train:9601-9700batch: iter_time=1.077e-04, loss_ctc=244.415, loss_att=194.236, acc=0.250, loss=209.368, grad_norm=6.797, loss_scale=1.000, learning_rate=4.109e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 18:51:52,982 (deepspeed_trainer:226) INFO: 2epoch:train:9701-9800batch: iter_time=1.092e-04, loss_ctc=260.805, loss_att=207.415, acc=0.240, loss=223.452, grad_norm=7.213, loss_scale=1.000, learning_rate=4.125e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 18:52:30,068 (deepspeed_trainer:226) INFO: 2epoch:train:9801-9900batch: iter_time=1.085e-04, loss_ctc=235.192, loss_att=186.961, acc=0.246, loss=201.383, grad_norm=6.268, loss_scale=1.000, learning_rate=4.142e-05, step_time=0.370 [2024-11-28 18:53:07,456] [INFO] [logging.py:129:log_dist] [Rank 0] step=25000, skipped=0, lr=[np.float64(4.1668333333333336e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:53:07,457] [INFO] [timer.py:264:stop] epoch=0/micro_step=25000/global_step=25000, RunningAvgSamplesPerSec=42.62456584125089, CurrSamplesPerSec=43.602151055297675, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:53:07,458 (deepspeed_trainer:226) INFO: 2epoch:train:9901-10000batch: iter_time=1.104e-04, loss_ctc=257.293, loss_att=201.357, acc=0.248, loss=218.115, grad_norm=6.526, loss_scale=1.000, learning_rate=4.159e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 18:53:44,525 (deepspeed_trainer:226) INFO: 2epoch:train:10001-10100batch: iter_time=1.090e-04, loss_ctc=263.007, loss_att=208.765, acc=0.248, loss=225.200, grad_norm=7.892, loss_scale=1.000, learning_rate=4.175e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 18:54:21,308 (deepspeed_trainer:226) INFO: 2epoch:train:10101-10200batch: iter_time=1.077e-04, loss_ctc=218.915, loss_att=175.755, acc=0.253, loss=188.877, grad_norm=6.161, loss_scale=1.000, learning_rate=4.192e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 18:54:58,468 (deepspeed_trainer:226) INFO: 2epoch:train:10201-10300batch: iter_time=1.069e-04, loss_ctc=270.875, loss_att=210.268, acc=0.243, loss=228.583, grad_norm=7.505, loss_scale=1.000, learning_rate=4.209e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 18:55:35,206 (deepspeed_trainer:226) INFO: 2epoch:train:10301-10400batch: iter_time=1.060e-04, loss_ctc=232.670, loss_att=187.335, acc=0.247, loss=200.965, grad_norm=6.656, loss_scale=1.000, learning_rate=4.225e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 18:56:12,496 (deepspeed_trainer:226) INFO: 2epoch:train:10401-10500batch: iter_time=1.089e-04, loss_ctc=230.485, loss_att=184.093, acc=0.252, loss=197.930, grad_norm=7.840, loss_scale=1.000, learning_rate=4.242e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:56:49,847 (deepspeed_trainer:226) INFO: 2epoch:train:10501-10600batch: iter_time=1.060e-04, loss_ctc=283.837, loss_att=214.438, acc=0.232, loss=235.350, grad_norm=7.792, loss_scale=1.000, learning_rate=4.259e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 18:57:26,971 (deepspeed_trainer:226) INFO: 2epoch:train:10601-10700batch: iter_time=1.043e-04, loss_ctc=240.400, loss_att=188.775, acc=0.243, loss=204.197, grad_norm=6.311, loss_scale=1.000, learning_rate=4.275e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 18:58:06,994 (deepspeed_trainer:226) INFO: 2epoch:train:10701-10800batch: iter_time=1.027e-04, loss_ctc=258.530, loss_att=205.295, acc=0.249, loss=221.322, grad_norm=7.499, loss_scale=1.000, learning_rate=4.292e-05, step_time=0.400 [cnode7-008:0/16] 2024-11-28 18:58:46,636 (deepspeed_trainer:226) INFO: 2epoch:train:10801-10900batch: iter_time=1.095e-04, loss_ctc=247.305, loss_att=195.180, acc=0.248, loss=210.838, grad_norm=7.637, loss_scale=1.000, learning_rate=4.309e-05, step_time=0.396 [2024-11-28 18:59:23,497] [INFO] [logging.py:129:log_dist] [Rank 0] step=26000, skipped=0, lr=[np.float64(4.3335e-05)], mom=[[0.9, 0.98]] [2024-11-28 18:59:23,498] [INFO] [timer.py:264:stop] epoch=0/micro_step=26000/global_step=26000, RunningAvgSamplesPerSec=42.65335856250865, CurrSamplesPerSec=41.939365212427404, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 18:59:23,499 (deepspeed_trainer:226) INFO: 2epoch:train:10901-11000batch: iter_time=1.083e-04, loss_ctc=236.625, loss_att=188.761, acc=0.246, loss=203.148, grad_norm=9.402, loss_scale=1.000, learning_rate=4.325e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:00:00,374 (deepspeed_trainer:226) INFO: 2epoch:train:11001-11100batch: iter_time=1.058e-04, loss_ctc=239.475, loss_att=193.188, acc=0.246, loss=207.072, grad_norm=7.463, loss_scale=1.000, learning_rate=4.342e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 19:00:37,232 (deepspeed_trainer:226) INFO: 2epoch:train:11101-11200batch: iter_time=1.083e-04, loss_ctc=255.325, loss_att=201.340, acc=0.243, loss=217.510, grad_norm=6.403, loss_scale=1.000, learning_rate=4.359e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:00:59,022 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-28 19:01:25,845 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 19:01:31,418 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 19:01:31,418 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-28 19:01:31,420 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 19:01:58,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:53,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:53,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:58,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:54,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:54,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:54,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:55,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:55,519] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:01:59,995] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:03,461] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:45,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:41,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:46,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:41,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:42,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:47,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:47,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:47,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:42,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:47,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:47,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:43,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:43,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:43,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:49,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:02:52,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:32,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:33,153] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:29,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:34,729] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:30,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:35,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:30,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:35,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:35,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:31,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:31,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:38,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:34,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:38,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:35,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:03:41,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:20,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:20,497] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:21,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:21,973] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:22,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:18,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:22,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:18,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:19,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:19,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:24,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:23,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:28,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:24,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:25,280] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:04:30,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 19:05:18,518 (deepspeed_trainer:226) INFO: 2epoch:train:11201-11300batch: iter_time=2.384, loss_ctc=250.697, loss_att=205.727, acc=0.247, loss=219.287, grad_norm=6.938, loss_scale=1.000, learning_rate=4.375e-05, step_time=0.429 [cnode7-008:0/16] 2024-11-28 19:05:55,683 (deepspeed_trainer:226) INFO: 2epoch:train:11301-11400batch: iter_time=1.044e-04, loss_ctc=239.822, loss_att=198.097, acc=0.241, loss=210.590, grad_norm=6.438, loss_scale=1.000, learning_rate=4.392e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 19:06:33,684 (deepspeed_trainer:226) INFO: 2epoch:train:11401-11500batch: iter_time=1.090e-04, loss_ctc=261.185, loss_att=217.282, acc=0.227, loss=230.450, grad_norm=6.882, loss_scale=1.000, learning_rate=4.409e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 19:07:11,574 (deepspeed_trainer:226) INFO: 2epoch:train:11501-11600batch: iter_time=1.048e-04, loss_ctc=251.834, loss_att=205.759, acc=0.250, loss=219.560, grad_norm=7.376, loss_scale=1.000, learning_rate=4.425e-05, step_time=0.378 [cnode7-008:0/16] 2024-11-28 19:07:48,860 (deepspeed_trainer:226) INFO: 2epoch:train:11601-11700batch: iter_time=1.059e-04, loss_ctc=228.792, loss_att=182.993, acc=0.243, loss=196.700, grad_norm=6.232, loss_scale=1.000, learning_rate=4.442e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 19:08:27,724 (deepspeed_trainer:226) INFO: 2epoch:train:11701-11800batch: iter_time=1.108e-04, loss_ctc=249.744, loss_att=203.385, acc=0.253, loss=217.278, grad_norm=7.049, loss_scale=1.000, learning_rate=4.459e-05, step_time=0.388 [cnode7-008:0/16] 2024-11-28 19:09:09,721 (deepspeed_trainer:226) INFO: 2epoch:train:11801-11900batch: iter_time=1.094e-04, loss_ctc=255.812, loss_att=199.620, acc=0.241, loss=216.440, grad_norm=7.809, loss_scale=1.000, learning_rate=4.475e-05, step_time=0.419 [2024-11-28 19:09:46,667] [INFO] [logging.py:129:log_dist] [Rank 0] step=27000, skipped=0, lr=[np.float64(4.5001666666666666e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:09:46,668] [INFO] [timer.py:264:stop] epoch=0/micro_step=27000/global_step=27000, RunningAvgSamplesPerSec=42.643253515133125, CurrSamplesPerSec=39.8266944060754, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:09:46,669 (deepspeed_trainer:226) INFO: 2epoch:train:11901-12000batch: iter_time=1.074e-04, loss_ctc=238.115, loss_att=197.070, acc=0.253, loss=209.438, grad_norm=7.199, loss_scale=1.000, learning_rate=4.492e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 19:10:23,732 (deepspeed_trainer:226) INFO: 2epoch:train:12001-12100batch: iter_time=1.049e-04, loss_ctc=220.252, loss_att=179.188, acc=0.256, loss=191.620, grad_norm=6.527, loss_scale=1.000, learning_rate=4.509e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 19:11:00,734 (deepspeed_trainer:226) INFO: 2epoch:train:12101-12200batch: iter_time=1.040e-04, loss_ctc=276.207, loss_att=221.463, acc=0.236, loss=237.990, grad_norm=9.013, loss_scale=1.000, learning_rate=4.525e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 19:11:37,525 (deepspeed_trainer:226) INFO: 2epoch:train:12201-12300batch: iter_time=1.059e-04, loss_ctc=221.855, loss_att=180.197, acc=0.255, loss=192.615, grad_norm=6.796, loss_scale=1.000, learning_rate=4.542e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 19:12:15,090 (deepspeed_trainer:226) INFO: 2epoch:train:12301-12400batch: iter_time=1.096e-04, loss_ctc=259.447, loss_att=206.084, acc=0.236, loss=222.120, grad_norm=9.233, loss_scale=1.000, learning_rate=4.559e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 19:12:52,308 (deepspeed_trainer:226) INFO: 2epoch:train:12401-12500batch: iter_time=1.047e-04, loss_ctc=252.523, loss_att=200.455, acc=0.246, loss=216.120, grad_norm=7.760, loss_scale=1.000, learning_rate=4.575e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 19:13:28,938 (deepspeed_trainer:226) INFO: 2epoch:train:12501-12600batch: iter_time=1.056e-04, loss_ctc=228.952, loss_att=187.425, acc=0.254, loss=199.863, grad_norm=6.672, loss_scale=1.000, learning_rate=4.592e-05, step_time=0.366 [cnode7-008:0/16] 2024-11-28 19:14:05,999 (deepspeed_trainer:226) INFO: 2epoch:train:12601-12700batch: iter_time=1.068e-04, loss_ctc=267.720, loss_att=218.062, acc=0.244, loss=232.950, grad_norm=6.888, loss_scale=1.000, learning_rate=4.609e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 19:14:42,832 (deepspeed_trainer:226) INFO: 2epoch:train:12701-12800batch: iter_time=1.048e-04, loss_ctc=224.322, loss_att=181.616, acc=0.251, loss=194.512, grad_norm=7.098, loss_scale=1.000, learning_rate=4.625e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:15:20,005 (deepspeed_trainer:226) INFO: 2epoch:train:12801-12900batch: iter_time=1.066e-04, loss_ctc=232.023, loss_att=193.541, acc=0.253, loss=205.115, grad_norm=7.715, loss_scale=1.000, learning_rate=4.642e-05, step_time=0.371 [2024-11-28 19:15:57,402] [INFO] [logging.py:129:log_dist] [Rank 0] step=28000, skipped=0, lr=[np.float64(4.6668333333333335e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:15:57,402] [INFO] [timer.py:264:stop] epoch=0/micro_step=28000/global_step=28000, RunningAvgSamplesPerSec=42.690664881177554, CurrSamplesPerSec=46.594473240532174, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:15:57,404 (deepspeed_trainer:226) INFO: 2epoch:train:12901-13000batch: iter_time=1.117e-04, loss_ctc=247.188, loss_att=202.137, acc=0.242, loss=215.665, grad_norm=7.183, loss_scale=1.000, learning_rate=4.659e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 19:16:34,797 (deepspeed_trainer:226) INFO: 2epoch:train:13001-13100batch: iter_time=1.103e-04, loss_ctc=245.002, loss_att=200.428, acc=0.250, loss=213.860, grad_norm=6.897, loss_scale=1.000, learning_rate=4.675e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 19:16:47,225 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-28 19:17:13,930 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 19:17:20,177 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 19:17:20,177 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-28 19:17:20,179 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 19:17:47,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:47,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:47,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:42,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:48,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:48,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:44,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:49,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:49,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:49,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:44,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:44,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:44,617] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:44,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:45,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:17:45,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:34,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:35,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:35,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:35,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:36,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:31,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:36,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:32,284] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:37,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:37,449] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:32,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:33,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:33,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:34,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:34,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:18:34,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:22,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:22,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:23,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:24,182] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:19,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:24,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:24,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:25,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:25,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:20,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:21,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:22,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:22,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:23,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:25,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:19:27,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:09,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:09,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:10,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:11,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:12,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:12,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:07,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:13,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:13,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:09,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:09,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:10,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:10,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:11,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:16,908] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:20:18,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 19:21:17,406 (deepspeed_trainer:226) INFO: 2epoch:train:13101-13200batch: iter_time=2.342, loss_ctc=245.835, loss_att=204.792, acc=0.253, loss=217.127, grad_norm=6.923, loss_scale=1.000, learning_rate=4.692e-05, step_time=0.483 [cnode7-008:0/16] 2024-11-28 19:21:54,818 (deepspeed_trainer:226) INFO: 2epoch:train:13201-13300batch: iter_time=1.069e-04, loss_ctc=239.020, loss_att=198.725, acc=0.238, loss=210.847, grad_norm=7.580, loss_scale=1.000, learning_rate=4.709e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 19:22:32,621 (deepspeed_trainer:226) INFO: 2epoch:train:13301-13400batch: iter_time=1.036e-04, loss_ctc=262.005, loss_att=219.680, acc=0.240, loss=232.360, grad_norm=8.417, loss_scale=1.000, learning_rate=4.725e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 19:23:10,199 (deepspeed_trainer:226) INFO: 2epoch:train:13401-13500batch: iter_time=1.064e-04, loss_ctc=228.338, loss_att=182.284, acc=0.254, loss=196.086, grad_norm=7.025, loss_scale=1.000, learning_rate=4.742e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 19:23:47,414 (deepspeed_trainer:226) INFO: 2epoch:train:13501-13600batch: iter_time=1.089e-04, loss_ctc=222.012, loss_att=177.720, acc=0.250, loss=190.977, grad_norm=6.701, loss_scale=1.000, learning_rate=4.759e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 19:24:25,133 (deepspeed_trainer:226) INFO: 2epoch:train:13601-13700batch: iter_time=1.060e-04, loss_ctc=257.999, loss_att=206.350, acc=0.251, loss=221.812, grad_norm=7.333, loss_scale=1.000, learning_rate=4.775e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 19:25:02,668 (deepspeed_trainer:226) INFO: 2epoch:train:13701-13800batch: iter_time=1.055e-04, loss_ctc=237.045, loss_att=192.434, acc=0.249, loss=205.785, grad_norm=6.766, loss_scale=1.000, learning_rate=4.792e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 19:25:39,950 (deepspeed_trainer:226) INFO: 2epoch:train:13801-13900batch: iter_time=1.058e-04, loss_ctc=230.455, loss_att=192.640, acc=0.255, loss=203.998, grad_norm=7.279, loss_scale=1.000, learning_rate=4.809e-05, step_time=0.372 [2024-11-28 19:26:17,871] [INFO] [logging.py:129:log_dist] [Rank 0] step=29000, skipped=0, lr=[np.float64(4.8335000000000004e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:26:17,871] [INFO] [timer.py:264:stop] epoch=0/micro_step=29000/global_step=29000, RunningAvgSamplesPerSec=42.67440192013703, CurrSamplesPerSec=42.569930265885674, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:26:17,873 (deepspeed_trainer:226) INFO: 2epoch:train:13901-14000batch: iter_time=1.065e-04, loss_ctc=234.572, loss_att=192.132, acc=0.255, loss=204.898, grad_norm=6.450, loss_scale=1.000, learning_rate=4.825e-05, step_time=0.379 [cnode7-008:0/16] 2024-11-28 19:26:55,958 (deepspeed_trainer:226) INFO: 2epoch:train:14001-14100batch: iter_time=1.070e-04, loss_ctc=257.200, loss_att=206.590, acc=0.243, loss=221.838, grad_norm=7.389, loss_scale=1.000, learning_rate=4.842e-05, step_time=0.381 [cnode7-008:0/16] 2024-11-28 19:27:32,974 (deepspeed_trainer:226) INFO: 2epoch:train:14101-14200batch: iter_time=1.066e-04, loss_ctc=212.222, loss_att=174.550, acc=0.263, loss=185.860, grad_norm=7.018, loss_scale=1.000, learning_rate=4.859e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 19:28:10,542 (deepspeed_trainer:226) INFO: 2epoch:train:14201-14300batch: iter_time=1.061e-04, loss_ctc=258.257, loss_att=201.857, acc=0.238, loss=218.790, grad_norm=8.469, loss_scale=1.000, learning_rate=4.875e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 19:28:49,694 (deepspeed_trainer:226) INFO: 2epoch:train:14301-14400batch: iter_time=1.048e-04, loss_ctc=246.760, loss_att=202.090, acc=0.252, loss=215.435, grad_norm=8.886, loss_scale=1.000, learning_rate=4.892e-05, step_time=0.392 [cnode7-008:0/16] 2024-11-28 19:29:30,079 (deepspeed_trainer:226) INFO: 2epoch:train:14401-14500batch: iter_time=1.028e-04, loss_ctc=227.667, loss_att=186.770, acc=0.254, loss=199.030, grad_norm=7.086, loss_scale=1.000, learning_rate=4.909e-05, step_time=0.403 [cnode7-008:0/16] 2024-11-28 19:30:07,532 (deepspeed_trainer:226) INFO: 2epoch:train:14501-14600batch: iter_time=1.044e-04, loss_ctc=251.502, loss_att=206.615, acc=0.254, loss=220.010, grad_norm=7.780, loss_scale=1.000, learning_rate=4.925e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 19:30:44,511 (deepspeed_trainer:226) INFO: 2epoch:train:14601-14700batch: iter_time=1.049e-04, loss_ctc=218.768, loss_att=179.373, acc=0.253, loss=191.248, grad_norm=7.083, loss_scale=1.000, learning_rate=4.942e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 19:31:21,957 (deepspeed_trainer:226) INFO: 2epoch:train:14701-14800batch: iter_time=1.058e-04, loss_ctc=228.482, loss_att=192.113, acc=0.252, loss=202.960, grad_norm=6.955, loss_scale=1.000, learning_rate=4.959e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 19:31:59,482 (deepspeed_trainer:226) INFO: 2epoch:train:14801-14900batch: iter_time=1.054e-04, loss_ctc=259.585, loss_att=214.243, acc=0.242, loss=227.934, grad_norm=7.276, loss_scale=1.000, learning_rate=4.975e-05, step_time=0.375 [2024-11-28 19:32:36,593] [INFO] [logging.py:129:log_dist] [Rank 0] step=30000, skipped=0, lr=[np.float64(5.0005e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:32:36,594] [INFO] [timer.py:264:stop] epoch=0/micro_step=30000/global_step=30000, RunningAvgSamplesPerSec=42.687863113781816, CurrSamplesPerSec=47.219188320684225, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:32:36,595 (deepspeed_trainer:226) INFO: 2epoch:train:14901-15000batch: iter_time=1.032e-04, loss_ctc=223.965, loss_att=182.553, acc=0.257, loss=194.917, grad_norm=8.042, loss_scale=1.000, learning_rate=4.992e-05, step_time=0.371 [2024-11-28 19:32:46,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:50,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:46,759] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:51,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:51,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:47,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:47,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:51,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:47,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:51,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:52,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:48,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:52,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:48,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:52,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:32:48,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:05,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:02,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:02,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:06,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:06,756] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:06,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:03,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:07,182] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:04,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:08,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:04,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:04,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:08,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:08,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:04,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:04,790] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:20,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:17,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:21,552] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:17,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:18,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:22,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:22,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:19,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:19,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:23,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:19,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:23,503] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:24,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:24,101] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:20,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:20,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:35,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:32,813] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:36,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:33,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:33,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:38,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:34,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:38,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:34,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:38,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:35,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:39,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:35,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:39,359] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:35,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:39,840] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:33:51,973] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 2 is about to be saved! [2024-11-28 19:33:52,005] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/mp_rank_00_model_states.pt [2024-11-28 19:33:52,005] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/mp_rank_00_model_states.pt... [2024-11-28 19:33:54,036] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/mp_rank_00_model_states.pt. [2024-11-28 19:33:54,194] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,289] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,289] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,291] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,197] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,292] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,198] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,198] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,200] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,201] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,202] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,296] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,296] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,297] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-28 19:33:50,297] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,203] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-28 19:33:54,951] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,955] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,956] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,070] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,070] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,070] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,071] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,071] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,071] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,979] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,979] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,979] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,073] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,073] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,073] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,981] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,981] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,981] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,981] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,981] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,981] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,982] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,982] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,983] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,997] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-28 19:33:54,997] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-28 19:33:54,997] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:54,999] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-28 19:33:55,000] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-28 19:33:55,000] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,099] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,100] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,100] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:55,008] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-28 19:33:55,008] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-28 19:33:55,008] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,104] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,104] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,104] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,122] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,122] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,122] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,166] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,166] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,166] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [2024-11-28 19:33:51,173] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-28 19:33:51,173] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_2/2/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-28 19:33:51,173] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 2 is ready now! [cnode7-008:0/16] 2024-11-28 19:33:55,089 (deepspeed_trainer:158) INFO: 2epoch results: [train] iter_time=0.125, loss_ctc=253.396, loss_att=205.356, acc=0.236, loss=219.778, grad_norm=7.916, loss_scale=1.000, learning_rate=3.750e-05, step_time=0.381, time=2 hours, 6 minutes and 47.42 seconds, total_count=30002, gpu_max_cached_mem_GB=27.396, [valid] loss_ctc=74.500, cer_ctc=0.995, loss_att=69.000, acc=0.241, cer=0.723, wer=1.000, loss=70.500, time=1 minute and 7.22 seconds, total_count=2, gpu_max_cached_mem_GB=27.396 [cnode7-008:0/16] 2024-11-28 19:33:57,075 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-28 19:34:23,675 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 19:34:29,281 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 19:34:29,282 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-28 19:34:29,284 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 19:34:47,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:47,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,613] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:47,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:51,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:48,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:48,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:48,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:53,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:49,359] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:49,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:34:54,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:34,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:34,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:38,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:38,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:35,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:35,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:39,594] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:39,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:35,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:39,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:39,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:35,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:40,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:41,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:37,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:35:37,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:25,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:25,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:21,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:26,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:26,578] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:22,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:22,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:27,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:23,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:27,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:23,613] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:23,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:27,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:25,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:25,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:36:29,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:11,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:13,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:13,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:13,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:10,049] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:14,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:10,825] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:10,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:14,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:15,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:11,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:11,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:11,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:12,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:13,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:37:18,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 19:38:21,994 (deepspeed_trainer:226) INFO: 3epoch:train:1-100batch: iter_time=2.229, loss_ctc=227.200, loss_att=187.616, acc=0.259, loss=199.465, grad_norm=6.467, loss_scale=1.000, learning_rate=5.026e-05, step_time=0.420 [cnode7-008:0/16] 2024-11-28 19:38:59,755 (deepspeed_trainer:226) INFO: 3epoch:train:101-200batch: iter_time=1.014e-04, loss_ctc=219.757, loss_att=174.806, acc=0.249, loss=188.278, grad_norm=7.332, loss_scale=1.000, learning_rate=5.076e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 19:39:36,905 (deepspeed_trainer:226) INFO: 3epoch:train:201-300batch: iter_time=1.069e-04, loss_ctc=246.905, loss_att=198.952, acc=0.246, loss=213.343, grad_norm=7.853, loss_scale=1.000, learning_rate=5.126e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 19:40:15,412 (deepspeed_trainer:226) INFO: 3epoch:train:301-400batch: iter_time=1.020e-04, loss_ctc=238.007, loss_att=190.536, acc=0.259, loss=204.715, grad_norm=7.733, loss_scale=1.000, learning_rate=5.176e-05, step_time=0.385 [cnode7-008:0/16] 2024-11-28 19:40:55,124 (deepspeed_trainer:226) INFO: 3epoch:train:401-500batch: iter_time=1.007e-04, loss_ctc=221.465, loss_att=176.847, acc=0.249, loss=190.287, grad_norm=9.038, loss_scale=1.000, learning_rate=5.226e-05, step_time=0.397 [cnode7-008:0/16] 2024-11-28 19:41:33,384 (deepspeed_trainer:226) INFO: 3epoch:train:501-600batch: iter_time=1.075e-04, loss_ctc=201.981, loss_att=157.889, acc=0.266, loss=171.124, grad_norm=8.137, loss_scale=1.000, learning_rate=5.276e-05, step_time=0.382 [cnode7-008:0/16] 2024-11-28 19:42:10,737 (deepspeed_trainer:226) INFO: 3epoch:train:601-700batch: iter_time=1.052e-04, loss_ctc=270.905, loss_att=214.010, acc=0.244, loss=230.965, grad_norm=10.625, loss_scale=1.000, learning_rate=5.326e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 19:42:47,586 (deepspeed_trainer:226) INFO: 3epoch:train:701-800batch: iter_time=1.039e-04, loss_ctc=229.577, loss_att=184.002, acc=0.250, loss=197.657, grad_norm=7.127, loss_scale=1.000, learning_rate=5.376e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:43:24,402 (deepspeed_trainer:226) INFO: 3epoch:train:801-900batch: iter_time=1.070e-04, loss_ctc=217.415, loss_att=178.727, acc=0.253, loss=190.338, grad_norm=6.646, loss_scale=1.000, learning_rate=5.426e-05, step_time=0.368 [2024-11-28 19:44:01,173] [INFO] [logging.py:129:log_dist] [Rank 0] step=31000, skipped=0, lr=[np.float64(5.5005000000000006e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:44:01,173] [INFO] [timer.py:264:stop] epoch=0/micro_step=31000/global_step=31000, RunningAvgSamplesPerSec=42.69018675509414, CurrSamplesPerSec=46.29659230597453, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:44:01,174 (deepspeed_trainer:226) INFO: 3epoch:train:901-1000batch: iter_time=1.078e-04, loss_ctc=210.648, loss_att=173.787, acc=0.259, loss=184.917, grad_norm=8.289, loss_scale=1.000, learning_rate=5.476e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:44:38,039 (deepspeed_trainer:226) INFO: 3epoch:train:1001-1100batch: iter_time=1.052e-04, loss_ctc=233.450, loss_att=187.398, acc=0.254, loss=201.132, grad_norm=7.251, loss_scale=1.000, learning_rate=5.526e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:45:15,032 (deepspeed_trainer:226) INFO: 3epoch:train:1101-1200batch: iter_time=1.048e-04, loss_ctc=233.083, loss_att=186.558, acc=0.246, loss=200.500, grad_norm=7.295, loss_scale=1.000, learning_rate=5.576e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 19:45:52,636 (deepspeed_trainer:226) INFO: 3epoch:train:1201-1300batch: iter_time=1.089e-04, loss_ctc=237.140, loss_att=194.560, acc=0.259, loss=207.230, grad_norm=6.192, loss_scale=1.000, learning_rate=5.626e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 19:46:30,273 (deepspeed_trainer:226) INFO: 3epoch:train:1301-1400batch: iter_time=1.056e-04, loss_ctc=240.520, loss_att=187.279, acc=0.253, loss=203.257, grad_norm=7.799, loss_scale=1.000, learning_rate=5.676e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 19:47:07,398 (deepspeed_trainer:226) INFO: 3epoch:train:1401-1500batch: iter_time=1.053e-04, loss_ctc=248.795, loss_att=193.905, acc=0.247, loss=210.345, grad_norm=8.046, loss_scale=1.000, learning_rate=5.726e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 19:47:45,517 (deepspeed_trainer:226) INFO: 3epoch:train:1501-1600batch: iter_time=1.069e-04, loss_ctc=234.055, loss_att=195.124, acc=0.256, loss=206.738, grad_norm=6.537, loss_scale=1.000, learning_rate=5.776e-05, step_time=0.381 [cnode7-008:0/16] 2024-11-28 19:48:23,965 (deepspeed_trainer:226) INFO: 3epoch:train:1601-1700batch: iter_time=1.056e-04, loss_ctc=220.762, loss_att=173.244, acc=0.266, loss=187.505, grad_norm=5.674, loss_scale=1.000, learning_rate=5.826e-05, step_time=0.384 [cnode7-008:0/16] 2024-11-28 19:49:03,676 (deepspeed_trainer:226) INFO: 3epoch:train:1701-1800batch: iter_time=1.059e-04, loss_ctc=238.045, loss_att=195.032, acc=0.244, loss=207.945, grad_norm=6.932, loss_scale=1.000, learning_rate=5.876e-05, step_time=0.396 [cnode7-008:0/16] 2024-11-28 19:49:36,378 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-28 19:50:03,874 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 19:50:09,649 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 19:50:09,649 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-28 19:50:09,652 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 19:50:34,409] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:34,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:38,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:38,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:34,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:34,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:34,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:34,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:35,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:35,033] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,473] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:50:39,945] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:22,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:23,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:23,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:23,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:28,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:28,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:28,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:28,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:24,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:24,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:24,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:24,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:29,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:30,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:30,488] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:51:30,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:11,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:11,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:11,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:16,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:16,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:12,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:17,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:12,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:17,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:13,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:13,743] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:18,037] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:18,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:18,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:14,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:19,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:58,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:59,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:52:59,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:04,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:00,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:05,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:05,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:01,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:05,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:01,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:01,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:07,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:07,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:08,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:08,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 19:53:05,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 19:53:45,739 (deepspeed_trainer:226) INFO: 3epoch:train:1801-1900batch: iter_time=2.387, loss_ctc=251.344, loss_att=203.553, acc=0.253, loss=217.865, grad_norm=6.343, loss_scale=1.000, learning_rate=5.926e-05, step_time=0.433 [2024-11-28 19:54:22,633] [INFO] [logging.py:129:log_dist] [Rank 0] step=32000, skipped=0, lr=[np.float64(6.0005000000000006e-05)], mom=[[0.9, 0.98]] [2024-11-28 19:54:22,634] [INFO] [timer.py:264:stop] epoch=0/micro_step=32000/global_step=32000, RunningAvgSamplesPerSec=42.686867757203686, CurrSamplesPerSec=44.96047234022908, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 19:54:22,635 (deepspeed_trainer:226) INFO: 3epoch:train:1901-2000batch: iter_time=1.021e-04, loss_ctc=197.007, loss_att=162.520, acc=0.267, loss=172.877, grad_norm=8.508, loss_scale=1.000, learning_rate=5.976e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 19:54:59,452 (deepspeed_trainer:226) INFO: 3epoch:train:2001-2100batch: iter_time=1.023e-04, loss_ctc=231.688, loss_att=190.856, acc=0.250, loss=203.070, grad_norm=6.666, loss_scale=1.000, learning_rate=6.026e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:55:37,002 (deepspeed_trainer:226) INFO: 3epoch:train:2101-2200batch: iter_time=1.071e-04, loss_ctc=230.982, loss_att=193.140, acc=0.253, loss=204.570, grad_norm=6.665, loss_scale=1.000, learning_rate=6.076e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 19:56:14,539 (deepspeed_trainer:226) INFO: 3epoch:train:2201-2300batch: iter_time=1.062e-04, loss_ctc=236.643, loss_att=197.162, acc=0.251, loss=208.988, grad_norm=6.270, loss_scale=1.000, learning_rate=6.126e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 19:56:51,324 (deepspeed_trainer:226) INFO: 3epoch:train:2301-2400batch: iter_time=1.052e-04, loss_ctc=222.602, loss_att=181.594, acc=0.250, loss=193.947, grad_norm=6.544, loss_scale=1.000, learning_rate=6.176e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 19:57:28,949 (deepspeed_trainer:226) INFO: 3epoch:train:2401-2500batch: iter_time=1.070e-04, loss_ctc=206.290, loss_att=165.386, acc=0.263, loss=177.732, grad_norm=6.178, loss_scale=1.000, learning_rate=6.226e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 19:58:08,305 (deepspeed_trainer:226) INFO: 3epoch:train:2501-2600batch: iter_time=1.096e-04, loss_ctc=250.270, loss_att=200.329, acc=0.252, loss=215.365, grad_norm=7.630, loss_scale=1.000, learning_rate=6.276e-05, step_time=0.389 [cnode7-008:0/16] 2024-11-28 19:58:48,148 (deepspeed_trainer:226) INFO: 3epoch:train:2601-2700batch: iter_time=1.042e-04, loss_ctc=233.085, loss_att=194.740, acc=0.246, loss=206.280, grad_norm=6.095, loss_scale=1.000, learning_rate=6.326e-05, step_time=0.402 [cnode7-008:0/16] 2024-11-28 19:59:25,101 (deepspeed_trainer:226) INFO: 3epoch:train:2701-2800batch: iter_time=1.048e-04, loss_ctc=202.450, loss_att=174.468, acc=0.262, loss=182.817, grad_norm=5.572, loss_scale=1.000, learning_rate=6.376e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:00:02,438 (deepspeed_trainer:226) INFO: 3epoch:train:2801-2900batch: iter_time=1.074e-04, loss_ctc=231.294, loss_att=193.405, acc=0.249, loss=204.739, grad_norm=7.546, loss_scale=1.000, learning_rate=6.426e-05, step_time=0.373 [2024-11-28 20:00:39,244] [INFO] [logging.py:129:log_dist] [Rank 0] step=33000, skipped=0, lr=[np.float64(6.5005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:00:39,245] [INFO] [timer.py:264:stop] epoch=0/micro_step=33000/global_step=33000, RunningAvgSamplesPerSec=42.705572940596724, CurrSamplesPerSec=43.59600446239575, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:00:39,246 (deepspeed_trainer:226) INFO: 3epoch:train:2901-3000batch: iter_time=1.068e-04, loss_ctc=204.655, loss_att=167.303, acc=0.263, loss=178.542, grad_norm=6.826, loss_scale=1.000, learning_rate=6.476e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:01:16,357 (deepspeed_trainer:226) INFO: 3epoch:train:3001-3100batch: iter_time=1.059e-04, loss_ctc=242.555, loss_att=200.190, acc=0.248, loss=212.850, grad_norm=6.174, loss_scale=1.000, learning_rate=6.526e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:01:53,368 (deepspeed_trainer:226) INFO: 3epoch:train:3101-3200batch: iter_time=1.053e-04, loss_ctc=224.512, loss_att=185.765, acc=0.261, loss=197.407, grad_norm=5.671, loss_scale=1.000, learning_rate=6.576e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 20:02:31,009 (deepspeed_trainer:226) INFO: 3epoch:train:3201-3300batch: iter_time=1.061e-04, loss_ctc=236.210, loss_att=188.407, acc=0.259, loss=202.725, grad_norm=7.142, loss_scale=1.000, learning_rate=6.626e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 20:03:08,131 (deepspeed_trainer:226) INFO: 3epoch:train:3301-3400batch: iter_time=1.070e-04, loss_ctc=255.440, loss_att=214.100, acc=0.243, loss=226.455, grad_norm=8.516, loss_scale=1.000, learning_rate=6.676e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 20:03:45,120 (deepspeed_trainer:226) INFO: 3epoch:train:3401-3500batch: iter_time=1.066e-04, loss_ctc=213.933, loss_att=182.090, acc=0.261, loss=191.519, grad_norm=7.423, loss_scale=1.000, learning_rate=6.726e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 20:04:22,007 (deepspeed_trainer:226) INFO: 3epoch:train:3501-3600batch: iter_time=1.064e-04, loss_ctc=224.822, loss_att=177.191, acc=0.259, loss=191.494, grad_norm=8.203, loss_scale=1.000, learning_rate=6.776e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:04:58,911 (deepspeed_trainer:226) INFO: 3epoch:train:3601-3700batch: iter_time=1.054e-04, loss_ctc=222.490, loss_att=194.511, acc=0.256, loss=202.916, grad_norm=5.220, loss_scale=1.000, learning_rate=6.826e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:05:21,269 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-28 20:05:47,037 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 20:05:52,499 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 20:05:52,499 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-28 20:05:52,501 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 20:06:20,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:20,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:20,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:20,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:21,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:22,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:22,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:22,475] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,833] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:06:19,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:08,698] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:09,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:09,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:09,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:06,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:07,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:07,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:11,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:11,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:11,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:11,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:07,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:07,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:08,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:08,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:08,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:55,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:56,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:57,169] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:58,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:54,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:59,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:59,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:55,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:59,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:00,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:56,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:56,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:57,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:57,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:57,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:07:57,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:42,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:44,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:44,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:41,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:46,550] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:44,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:44,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:48,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:44,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:48,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:48,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:45,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:49,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:45,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:45,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:08:46,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 20:09:36,176 (deepspeed_trainer:226) INFO: 3epoch:train:3701-3800batch: iter_time=2.303, loss_ctc=246.374, loss_att=202.375, acc=0.255, loss=215.555, grad_norm=6.201, loss_scale=1.000, learning_rate=6.876e-05, step_time=0.469 [cnode7-008:0/16] 2024-11-28 20:10:13,286 (deepspeed_trainer:226) INFO: 3epoch:train:3801-3900batch: iter_time=1.180e-04, loss_ctc=195.468, loss_att=162.546, acc=0.268, loss=172.398, grad_norm=5.004, loss_scale=1.000, learning_rate=6.926e-05, step_time=0.371 [2024-11-28 20:10:51,212] [INFO] [logging.py:129:log_dist] [Rank 0] step=34000, skipped=0, lr=[np.float64(7.0005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:10:51,213] [INFO] [timer.py:264:stop] epoch=0/micro_step=34000/global_step=34000, RunningAvgSamplesPerSec=42.70585689785288, CurrSamplesPerSec=44.49112473121327, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:10:51,214 (deepspeed_trainer:226) INFO: 3epoch:train:3901-4000batch: iter_time=1.020e-04, loss_ctc=220.817, loss_att=180.939, acc=0.253, loss=192.968, grad_norm=5.601, loss_scale=1.000, learning_rate=6.976e-05, step_time=0.379 [cnode7-008:0/16] 2024-11-28 20:11:28,455 (deepspeed_trainer:226) INFO: 3epoch:train:4001-4100batch: iter_time=1.034e-04, loss_ctc=236.773, loss_att=203.007, acc=0.251, loss=213.121, grad_norm=5.392, loss_scale=1.000, learning_rate=7.026e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 20:12:05,265 (deepspeed_trainer:226) INFO: 3epoch:train:4101-4200batch: iter_time=1.140e-04, loss_ctc=225.754, loss_att=185.929, acc=0.259, loss=197.934, grad_norm=6.229, loss_scale=1.000, learning_rate=7.076e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:12:41,706 (deepspeed_trainer:226) INFO: 3epoch:train:4201-4300batch: iter_time=1.053e-04, loss_ctc=202.007, loss_att=163.530, acc=0.264, loss=175.069, grad_norm=6.363, loss_scale=1.000, learning_rate=7.126e-05, step_time=0.364 [cnode7-008:0/16] 2024-11-28 20:13:18,769 (deepspeed_trainer:226) INFO: 3epoch:train:4301-4400batch: iter_time=1.040e-04, loss_ctc=239.468, loss_att=193.484, acc=0.254, loss=207.363, grad_norm=7.137, loss_scale=1.000, learning_rate=7.176e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 20:13:55,973 (deepspeed_trainer:226) INFO: 3epoch:train:4401-4500batch: iter_time=1.043e-04, loss_ctc=229.657, loss_att=186.990, acc=0.259, loss=199.702, grad_norm=8.149, loss_scale=1.000, learning_rate=7.226e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:14:32,830 (deepspeed_trainer:226) INFO: 3epoch:train:4501-4600batch: iter_time=1.069e-04, loss_ctc=218.868, loss_att=185.428, acc=0.252, loss=195.493, grad_norm=6.112, loss_scale=1.000, learning_rate=7.276e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:15:10,092 (deepspeed_trainer:226) INFO: 3epoch:train:4601-4700batch: iter_time=1.072e-04, loss_ctc=199.945, loss_att=174.192, acc=0.267, loss=181.903, grad_norm=5.642, loss_scale=1.000, learning_rate=7.326e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 20:15:47,510 (deepspeed_trainer:226) INFO: 3epoch:train:4701-4800batch: iter_time=1.062e-04, loss_ctc=217.842, loss_att=181.418, acc=0.257, loss=192.347, grad_norm=5.943, loss_scale=1.000, learning_rate=7.376e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 20:16:26,127 (deepspeed_trainer:226) INFO: 3epoch:train:4801-4900batch: iter_time=1.043e-04, loss_ctc=214.368, loss_att=175.347, acc=0.256, loss=187.083, grad_norm=5.952, loss_scale=1.000, learning_rate=7.426e-05, step_time=0.386 [2024-11-28 20:17:04,192] [INFO] [logging.py:129:log_dist] [Rank 0] step=35000, skipped=0, lr=[np.float64(7.5005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:17:04,192] [INFO] [timer.py:264:stop] epoch=0/micro_step=35000/global_step=35000, RunningAvgSamplesPerSec=42.73471633517278, CurrSamplesPerSec=42.14617946679793, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:17:04,194 (deepspeed_trainer:226) INFO: 3epoch:train:4901-5000batch: iter_time=1.052e-04, loss_ctc=228.088, loss_att=194.081, acc=0.259, loss=204.243, grad_norm=5.195, loss_scale=1.000, learning_rate=7.476e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 20:17:41,578 (deepspeed_trainer:226) INFO: 3epoch:train:5001-5100batch: iter_time=1.034e-04, loss_ctc=220.560, loss_att=183.256, acc=0.259, loss=194.447, grad_norm=5.448, loss_scale=1.000, learning_rate=7.526e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 20:18:18,202 (deepspeed_trainer:226) INFO: 3epoch:train:5101-5200batch: iter_time=1.051e-04, loss_ctc=245.363, loss_att=197.839, acc=0.258, loss=212.113, grad_norm=6.040, loss_scale=1.000, learning_rate=7.576e-05, step_time=0.366 [cnode7-008:0/16] 2024-11-28 20:18:55,003 (deepspeed_trainer:226) INFO: 3epoch:train:5201-5300batch: iter_time=1.060e-04, loss_ctc=237.308, loss_att=201.219, acc=0.250, loss=212.105, grad_norm=6.040, loss_scale=1.000, learning_rate=7.626e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 20:19:31,781 (deepspeed_trainer:226) INFO: 3epoch:train:5301-5400batch: iter_time=1.034e-04, loss_ctc=210.800, loss_att=177.945, acc=0.265, loss=187.751, grad_norm=5.315, loss_scale=1.000, learning_rate=7.676e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 20:20:08,337 (deepspeed_trainer:226) INFO: 3epoch:train:5401-5500batch: iter_time=1.052e-04, loss_ctc=212.287, loss_att=174.248, acc=0.264, loss=185.836, grad_norm=5.888, loss_scale=1.000, learning_rate=7.726e-05, step_time=0.365 [cnode7-008:0/16] 2024-11-28 20:20:45,251 (deepspeed_trainer:226) INFO: 3epoch:train:5501-5600batch: iter_time=1.068e-04, loss_ctc=228.007, loss_att=195.015, acc=0.254, loss=204.890, grad_norm=4.747, loss_scale=1.000, learning_rate=7.776e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:20:57,819 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-28 20:21:24,463 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 20:21:29,971 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 20:21:29,971 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-28 20:21:29,974 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 20:21:53,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:53,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:57,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:58,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:58,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:58,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:54,144] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:54,361] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:58,881] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:59,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:59,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:55,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:55,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:55,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:21:55,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:00,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:46,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:42,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:46,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:47,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:42,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:43,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:48,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:48,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:43,955] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:44,169] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:48,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:44,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:44,495] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:44,497] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:48,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:22:49,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:34,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:34,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:30,824] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:35,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:35,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:35,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:31,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:36,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:32,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:32,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:32,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:37,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:32,855] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:37,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:35,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:23:36,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:21,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:22,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:18,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:23,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:19,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:24,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:24,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:24,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:24,833] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:20,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:25,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:20,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:21,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:23,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:25,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:24:27,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 20:25:25,829 (deepspeed_trainer:226) INFO: 3epoch:train:5601-5700batch: iter_time=2.342, loss_ctc=231.149, loss_att=193.692, acc=0.268, loss=204.955, grad_norm=5.188, loss_scale=1.000, learning_rate=7.826e-05, step_time=0.463 [cnode7-008:0/16] 2024-11-28 20:26:04,725 (deepspeed_trainer:226) INFO: 3epoch:train:5701-5800batch: iter_time=1.042e-04, loss_ctc=196.625, loss_att=161.227, acc=0.261, loss=171.845, grad_norm=5.216, loss_scale=1.000, learning_rate=7.876e-05, step_time=0.388 [cnode7-008:0/16] 2024-11-28 20:26:45,028 (deepspeed_trainer:226) INFO: 3epoch:train:5801-5900batch: iter_time=1.028e-04, loss_ctc=219.335, loss_att=181.118, acc=0.259, loss=192.562, grad_norm=5.421, loss_scale=1.000, learning_rate=7.926e-05, step_time=0.403 [2024-11-28 20:27:22,607] [INFO] [logging.py:129:log_dist] [Rank 0] step=36000, skipped=0, lr=[np.float64(8.0005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:27:22,607] [INFO] [timer.py:264:stop] epoch=0/micro_step=36000/global_step=36000, RunningAvgSamplesPerSec=42.726924664985575, CurrSamplesPerSec=44.1025432379215, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:27:22,609 (deepspeed_trainer:226) INFO: 3epoch:train:5901-6000batch: iter_time=1.073e-04, loss_ctc=231.053, loss_att=194.799, acc=0.265, loss=205.725, grad_norm=4.817, loss_scale=1.000, learning_rate=7.976e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 20:27:59,799 (deepspeed_trainer:226) INFO: 3epoch:train:6001-6100batch: iter_time=1.031e-04, loss_ctc=218.074, loss_att=177.667, acc=0.258, loss=189.749, grad_norm=5.230, loss_scale=1.000, learning_rate=8.026e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:28:36,724 (deepspeed_trainer:226) INFO: 3epoch:train:6101-6200batch: iter_time=1.047e-04, loss_ctc=191.548, loss_att=155.632, acc=0.269, loss=166.426, grad_norm=4.741, loss_scale=1.000, learning_rate=8.076e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:29:14,023 (deepspeed_trainer:226) INFO: 3epoch:train:6201-6300batch: iter_time=1.067e-04, loss_ctc=240.338, loss_att=191.924, acc=0.255, loss=206.584, grad_norm=6.630, loss_scale=1.000, learning_rate=8.126e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 20:29:51,166 (deepspeed_trainer:226) INFO: 3epoch:train:6301-6400batch: iter_time=1.055e-04, loss_ctc=232.202, loss_att=190.260, acc=0.256, loss=202.787, grad_norm=5.669, loss_scale=1.000, learning_rate=8.176e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:30:28,039 (deepspeed_trainer:226) INFO: 3epoch:train:6401-6500batch: iter_time=1.054e-04, loss_ctc=206.685, loss_att=175.186, acc=0.261, loss=184.570, grad_norm=5.007, loss_scale=1.000, learning_rate=8.226e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:31:04,829 (deepspeed_trainer:226) INFO: 3epoch:train:6501-6600batch: iter_time=1.087e-04, loss_ctc=194.960, loss_att=167.628, acc=0.272, loss=175.897, grad_norm=4.157, loss_scale=1.000, learning_rate=8.276e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:31:42,473 (deepspeed_trainer:226) INFO: 3epoch:train:6601-6700batch: iter_time=1.076e-04, loss_ctc=216.320, loss_att=175.845, acc=0.260, loss=187.955, grad_norm=5.194, loss_scale=1.000, learning_rate=8.326e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 20:32:19,857 (deepspeed_trainer:226) INFO: 3epoch:train:6701-6800batch: iter_time=1.071e-04, loss_ctc=203.665, loss_att=167.190, acc=0.263, loss=178.137, grad_norm=5.122, loss_scale=1.000, learning_rate=8.376e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 20:32:56,966 (deepspeed_trainer:226) INFO: 3epoch:train:6801-6900batch: iter_time=1.071e-04, loss_ctc=230.540, loss_att=190.498, acc=0.259, loss=202.482, grad_norm=4.831, loss_scale=1.000, learning_rate=8.426e-05, step_time=0.371 [2024-11-28 20:33:34,534] [INFO] [logging.py:129:log_dist] [Rank 0] step=37000, skipped=0, lr=[np.float64(8.5005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:33:34,535] [INFO] [timer.py:264:stop] epoch=0/micro_step=37000/global_step=37000, RunningAvgSamplesPerSec=42.75720090355153, CurrSamplesPerSec=42.21322590266912, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:33:34,536 (deepspeed_trainer:226) INFO: 3epoch:train:6901-7000batch: iter_time=1.080e-04, loss_ctc=210.475, loss_att=173.790, acc=0.271, loss=184.850, grad_norm=4.864, loss_scale=1.000, learning_rate=8.476e-05, step_time=0.375 [cnode7-008:0/16] 2024-11-28 20:34:13,094 (deepspeed_trainer:226) INFO: 3epoch:train:7001-7100batch: iter_time=1.090e-04, loss_ctc=237.690, loss_att=187.752, acc=0.260, loss=202.697, grad_norm=5.631, loss_scale=1.000, learning_rate=8.526e-05, step_time=0.385 [cnode7-008:0/16] 2024-11-28 20:34:53,131 (deepspeed_trainer:226) INFO: 3epoch:train:7101-7200batch: iter_time=1.068e-04, loss_ctc=229.680, loss_att=194.896, acc=0.255, loss=205.471, grad_norm=5.565, loss_scale=1.000, learning_rate=8.576e-05, step_time=0.400 [cnode7-008:0/16] 2024-11-28 20:35:30,996 (deepspeed_trainer:226) INFO: 3epoch:train:7201-7300batch: iter_time=1.072e-04, loss_ctc=200.576, loss_att=164.514, acc=0.276, loss=175.442, grad_norm=5.548, loss_scale=1.000, learning_rate=8.626e-05, step_time=0.378 [cnode7-008:0/16] 2024-11-28 20:36:08,193 (deepspeed_trainer:226) INFO: 3epoch:train:7301-7400batch: iter_time=1.066e-04, loss_ctc=221.782, loss_att=184.403, acc=0.258, loss=195.558, grad_norm=4.488, loss_scale=1.000, learning_rate=8.676e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 20:36:45,976 (deepspeed_trainer:226) INFO: 3epoch:train:7401-7500batch: iter_time=1.055e-04, loss_ctc=225.235, loss_att=188.931, acc=0.261, loss=199.782, grad_norm=5.509, loss_scale=1.000, learning_rate=8.726e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 20:36:49,128 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-28 20:37:15,637 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 20:37:21,171 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 20:37:21,171 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-28 20:37:21,173 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 20:37:43,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:48,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:48,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:48,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:44,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:49,249] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:49,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:44,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:49,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:45,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:50,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:45,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:50,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:46,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:46,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:37:46,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:36,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:37,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:37,461] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:33,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:37,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:38,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:38,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:34,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:34,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:34,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:34,613] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:34,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:39,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:39,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:35,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:38:37,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:24,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:24,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:25,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:20,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:25,540] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:21,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:26,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:22,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:22,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:22,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:27,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:27,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:23,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:28,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:27,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:39:27,784] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:12,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:12,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:08,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:14,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:14,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:10,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:15,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:15,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:11,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:11,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:11,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:17,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:12,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:19,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:15,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:40:16,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 20:41:25,572 (deepspeed_trainer:226) INFO: 3epoch:train:7501-7600batch: iter_time=2.335, loss_ctc=209.965, loss_att=182.981, acc=0.269, loss=191.044, grad_norm=4.691, loss_scale=1.000, learning_rate=8.776e-05, step_time=0.461 [cnode7-008:0/16] 2024-11-28 20:42:03,470 (deepspeed_trainer:226) INFO: 3epoch:train:7601-7700batch: iter_time=1.071e-04, loss_ctc=203.090, loss_att=169.102, acc=0.261, loss=179.317, grad_norm=4.465, loss_scale=1.000, learning_rate=8.826e-05, step_time=0.378 [cnode7-008:0/16] 2024-11-28 20:42:40,777 (deepspeed_trainer:226) INFO: 3epoch:train:7701-7800batch: iter_time=1.063e-04, loss_ctc=225.375, loss_att=193.843, acc=0.254, loss=203.350, grad_norm=4.759, loss_scale=1.000, learning_rate=8.876e-05, step_time=0.373 [cnode7-008:0/16] 2024-11-28 20:43:17,798 (deepspeed_trainer:226) INFO: 3epoch:train:7801-7900batch: iter_time=1.087e-04, loss_ctc=220.488, loss_att=189.025, acc=0.263, loss=198.449, grad_norm=4.979, loss_scale=1.000, learning_rate=8.926e-05, step_time=0.370 [2024-11-28 20:43:55,087] [INFO] [logging.py:129:log_dist] [Rank 0] step=38000, skipped=0, lr=[np.float64(9.0005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:43:55,088] [INFO] [timer.py:264:stop] epoch=0/micro_step=38000/global_step=38000, RunningAvgSamplesPerSec=42.740431168289945, CurrSamplesPerSec=41.42631931014247, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:43:55,089 (deepspeed_trainer:226) INFO: 3epoch:train:7901-8000batch: iter_time=1.068e-04, loss_ctc=203.603, loss_att=171.322, acc=0.260, loss=181.005, grad_norm=4.465, loss_scale=1.000, learning_rate=8.976e-05, step_time=0.372 [cnode7-008:0/16] 2024-11-28 20:44:34,197 (deepspeed_trainer:226) INFO: 3epoch:train:8001-8100batch: iter_time=1.071e-04, loss_ctc=185.096, loss_att=154.161, acc=0.274, loss=163.446, grad_norm=4.171, loss_scale=1.000, learning_rate=9.026e-05, step_time=0.391 [cnode7-008:0/16] 2024-11-28 20:45:13,177 (deepspeed_trainer:226) INFO: 3epoch:train:8101-8200batch: iter_time=1.089e-04, loss_ctc=246.330, loss_att=203.666, acc=0.254, loss=216.407, grad_norm=6.363, loss_scale=1.000, learning_rate=9.076e-05, step_time=0.390 [cnode7-008:0/16] 2024-11-28 20:45:51,199 (deepspeed_trainer:226) INFO: 3epoch:train:8201-8300batch: iter_time=1.067e-04, loss_ctc=210.515, loss_att=180.493, acc=0.262, loss=189.410, grad_norm=4.324, loss_scale=1.000, learning_rate=9.126e-05, step_time=0.380 [cnode7-008:0/16] 2024-11-28 20:46:28,172 (deepspeed_trainer:226) INFO: 3epoch:train:8301-8400batch: iter_time=1.071e-04, loss_ctc=197.850, loss_att=171.643, acc=0.263, loss=179.512, grad_norm=3.733, loss_scale=1.000, learning_rate=9.176e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:47:05,131 (deepspeed_trainer:226) INFO: 3epoch:train:8401-8500batch: iter_time=1.095e-04, loss_ctc=193.945, loss_att=169.612, acc=0.271, loss=176.890, grad_norm=3.967, loss_scale=1.000, learning_rate=9.226e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:47:42,886 (deepspeed_trainer:226) INFO: 3epoch:train:8501-8600batch: iter_time=1.051e-04, loss_ctc=212.769, loss_att=183.951, acc=0.262, loss=192.655, grad_norm=4.705, loss_scale=1.000, learning_rate=9.276e-05, step_time=0.377 [cnode7-008:0/16] 2024-11-28 20:48:20,048 (deepspeed_trainer:226) INFO: 3epoch:train:8601-8700batch: iter_time=1.075e-04, loss_ctc=213.077, loss_att=181.178, acc=0.258, loss=190.792, grad_norm=4.857, loss_scale=1.000, learning_rate=9.326e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:48:57,437 (deepspeed_trainer:226) INFO: 3epoch:train:8701-8800batch: iter_time=1.072e-04, loss_ctc=215.144, loss_att=188.759, acc=0.269, loss=196.685, grad_norm=3.936, loss_scale=1.000, learning_rate=9.376e-05, step_time=0.374 [cnode7-008:0/16] 2024-11-28 20:49:34,520 (deepspeed_trainer:226) INFO: 3epoch:train:8801-8900batch: iter_time=1.076e-04, loss_ctc=217.192, loss_att=177.181, acc=0.266, loss=189.256, grad_norm=5.079, loss_scale=1.000, learning_rate=9.426e-05, step_time=0.370 [2024-11-28 20:50:11,641] [INFO] [logging.py:129:log_dist] [Rank 0] step=39000, skipped=0, lr=[np.float64(9.5005e-05)], mom=[[0.9, 0.98]] [2024-11-28 20:50:11,641] [INFO] [timer.py:264:stop] epoch=0/micro_step=39000/global_step=39000, RunningAvgSamplesPerSec=42.75523650281232, CurrSamplesPerSec=44.978763750188534, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 20:50:11,643 (deepspeed_trainer:226) INFO: 3epoch:train:8901-9000batch: iter_time=1.072e-04, loss_ctc=225.900, loss_att=190.827, acc=0.256, loss=201.425, grad_norm=5.191, loss_scale=1.000, learning_rate=9.476e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:50:49,824 (deepspeed_trainer:226) INFO: 3epoch:train:9001-9100batch: iter_time=1.088e-04, loss_ctc=214.204, loss_att=190.220, acc=0.265, loss=197.477, grad_norm=4.871, loss_scale=1.000, learning_rate=9.526e-05, step_time=0.382 [cnode7-008:0/16] 2024-11-28 20:51:26,839 (deepspeed_trainer:226) INFO: 3epoch:train:9101-9200batch: iter_time=1.085e-04, loss_ctc=200.891, loss_att=169.739, acc=0.275, loss=179.075, grad_norm=4.085, loss_scale=1.000, learning_rate=9.576e-05, step_time=0.370 [cnode7-008:0/16] 2024-11-28 20:52:04,481 (deepspeed_trainer:226) INFO: 3epoch:train:9201-9300batch: iter_time=1.073e-04, loss_ctc=216.738, loss_att=190.006, acc=0.258, loss=198.007, grad_norm=3.872, loss_scale=1.000, learning_rate=9.626e-05, step_time=0.376 [cnode7-008:0/16] 2024-11-28 20:52:35,460 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-28 20:53:02,058 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 20:53:07,561 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 20:53:07,561 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-28 20:53:07,563 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 20:53:34,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:35,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:30,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:36,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:36,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:36,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:31,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:31,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:37,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:37,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:31,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:38,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:32,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:32,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:32,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:53:32,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:22,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:24,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:19,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:25,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:25,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:19,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:25,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:20,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:20,774] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:20,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:21,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:27,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:22,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:27,801] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:27,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:54:24,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:09,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:11,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:06,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:12,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:07,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:13,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:13,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:08,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:09,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:10,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:10,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:10,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:16,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:17,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:17,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:13,973] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:57,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:59,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:59,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:54,610] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:01,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:01,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:55,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:56,173] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:57,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:57,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:58,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:55:58,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:04,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:05,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:05,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 20:56:03,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 20:56:43,330 (deepspeed_trainer:226) INFO: 3epoch:train:9301-9400batch: iter_time=2.329, loss_ctc=228.849, loss_att=195.333, acc=0.265, loss=205.379, grad_norm=5.203, loss_scale=1.000, learning_rate=9.676e-05, step_time=0.459 [cnode7-008:0/16] 2024-11-28 20:57:20,250 (deepspeed_trainer:226) INFO: 3epoch:train:9401-9500batch: iter_time=1.057e-04, loss_ctc=181.118, loss_att=152.944, acc=0.282, loss=161.386, grad_norm=4.247, loss_scale=1.000, learning_rate=9.726e-05, step_time=0.369 [cnode7-008:0/16] 2024-11-28 20:57:57,385 (deepspeed_trainer:226) INFO: 3epoch:train:9501-9600batch: iter_time=1.045e-04, loss_ctc=210.532, loss_att=179.748, acc=0.261, loss=188.969, grad_norm=3.780, loss_scale=1.000, learning_rate=9.776e-05, step_time=0.371 [cnode7-008:0/16] 2024-11-28 20:58:34,158 (deepspeed_trainer:226) INFO: 3epoch:train:9601-9700batch: iter_time=1.024e-04, loss_ctc=208.162, loss_att=176.974, acc=0.272, loss=186.357, grad_norm=4.495, loss_scale=1.000, learning_rate=9.826e-05, step_time=0.367 [cnode7-008:0/16] 2024-11-28 20:59:10,974 (deepspeed_trainer:226) INFO: 3epoch:train:9701-9800batch: iter_time=1.067e-04, loss_ctc=217.553, loss_att=182.811, acc=0.264, loss=193.222, grad_norm=4.952, loss_scale=1.000, learning_rate=9.876e-05, step_time=0.368 [cnode7-008:0/16] 2024-11-28 20:59:47,572 (deepspeed_trainer:226) INFO: 3epoch:train:9801-9900batch: iter_time=1.070e-04, loss_ctc=203.792, loss_att=171.717, acc=0.261, loss=181.393, grad_norm=4.531, loss_scale=1.000, learning_rate=9.926e-05, step_time=0.366 [2024-11-28 21:00:24,163] [INFO] [logging.py:129:log_dist] [Rank 0] step=40000, skipped=0, lr=[np.float64(0.000100005)], mom=[[0.9, 0.98]] [2024-11-28 21:00:24,164] [INFO] [timer.py:264:stop] epoch=0/micro_step=40000/global_step=40000, RunningAvgSamplesPerSec=42.76056298924416, CurrSamplesPerSec=46.689195297260035, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:00:24,165 (deepspeed_trainer:226) INFO: 3epoch:train:9901-10000batch: iter_time=1.087e-04, loss_ctc=187.341, loss_att=153.495, acc=0.276, loss=163.734, grad_norm=5.627, loss_scale=1.000, learning_rate=9.976e-05, step_time=0.366 [cnode7-008:0/16] 2024-11-28 21:01:02,166 (deepspeed_trainer:226) INFO: 3epoch:train:10001-10100batch: iter_time=1.070e-04, loss_ctc=228.266, loss_att=187.174, acc=0.266, loss=199.515, grad_norm=6.021, loss_scale=1.000, learning_rate=1.003e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-28 21:01:38,827 (deepspeed_trainer:226) INFO: 3epoch:train:10101-10200batch: iter_time=1.081e-04, loss_ctc=212.050, loss_att=180.675, acc=0.260, loss=190.148, grad_norm=4.357, loss_scale=1.000, learning_rate=1.008e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 21:02:15,905 (deepspeed_trainer:226) INFO: 3epoch:train:10201-10300batch: iter_time=1.067e-04, loss_ctc=184.701, loss_att=164.380, acc=0.275, loss=170.380, grad_norm=3.417, loss_scale=1.000, learning_rate=1.013e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 21:02:52,623 (deepspeed_trainer:226) INFO: 3epoch:train:10301-10400batch: iter_time=1.067e-04, loss_ctc=211.518, loss_att=179.282, acc=0.265, loss=189.012, grad_norm=5.075, loss_scale=1.000, learning_rate=1.018e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 21:03:30,382 (deepspeed_trainer:226) INFO: 3epoch:train:10401-10500batch: iter_time=1.075e-04, loss_ctc=186.037, loss_att=156.324, acc=0.276, loss=165.259, grad_norm=4.173, loss_scale=1.000, learning_rate=1.023e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 21:04:09,226 (deepspeed_trainer:226) INFO: 3epoch:train:10501-10600batch: iter_time=1.070e-04, loss_ctc=218.535, loss_att=186.655, acc=0.259, loss=196.218, grad_norm=4.389, loss_scale=1.000, learning_rate=1.028e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-28 21:04:47,519 (deepspeed_trainer:226) INFO: 3epoch:train:10601-10700batch: iter_time=1.074e-04, loss_ctc=203.999, loss_att=174.610, acc=0.273, loss=183.360, grad_norm=3.689, loss_scale=1.000, learning_rate=1.033e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-28 21:05:24,100 (deepspeed_trainer:226) INFO: 3epoch:train:10701-10800batch: iter_time=1.071e-04, loss_ctc=214.363, loss_att=173.118, acc=0.276, loss=185.488, grad_norm=4.554, loss_scale=1.000, learning_rate=1.038e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-28 21:06:00,842 (deepspeed_trainer:226) INFO: 3epoch:train:10801-10900batch: iter_time=1.071e-04, loss_ctc=232.333, loss_att=199.721, acc=0.258, loss=209.491, grad_norm=5.320, loss_scale=1.000, learning_rate=1.043e-04, step_time=0.366 [2024-11-28 21:06:37,346] [INFO] [logging.py:129:log_dist] [Rank 0] step=41000, skipped=0, lr=[np.float64(0.000105005)], mom=[[0.9, 0.98]] [2024-11-28 21:06:37,346] [INFO] [timer.py:264:stop] epoch=0/micro_step=41000/global_step=41000, RunningAvgSamplesPerSec=42.78378616414991, CurrSamplesPerSec=45.059943467420624, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:06:37,348 (deepspeed_trainer:226) INFO: 3epoch:train:10901-11000batch: iter_time=1.081e-04, loss_ctc=193.569, loss_att=167.856, acc=0.279, loss=175.581, grad_norm=4.139, loss_scale=1.000, learning_rate=1.048e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 21:07:14,219 (deepspeed_trainer:226) INFO: 3epoch:train:11001-11100batch: iter_time=1.038e-04, loss_ctc=203.610, loss_att=164.532, acc=0.273, loss=176.252, grad_norm=4.301, loss_scale=1.000, learning_rate=1.053e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 21:07:50,916 (deepspeed_trainer:226) INFO: 3epoch:train:11101-11200batch: iter_time=1.072e-04, loss_ctc=203.205, loss_att=182.970, acc=0.267, loss=189.119, grad_norm=3.492, loss_scale=1.000, learning_rate=1.058e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 21:08:12,518 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-28 21:08:39,518 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 21:08:45,011 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 21:08:45,011 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-28 21:08:45,013 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 21:09:12,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:12,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:12,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:07,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:07,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:13,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:08,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:13,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:13,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:08,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:14,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:14,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:09,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:09,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:09,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:09,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:54,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:00,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:55,704] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:01,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:56,108] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:56,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:56,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,855] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:02,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:57,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:57,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:09:57,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:42,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:43,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:49,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:44,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:50,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:44,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:44,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:50,529] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:45,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:45,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:51,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:45,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:51,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:52,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:53,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:10:54,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:29,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:36,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:31,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:32,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:38,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:33,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:33,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:39,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:39,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:34,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:34,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:34,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:39,881] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:40,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:43,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:11:43,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 21:12:28,565 (deepspeed_trainer:226) INFO: 3epoch:train:11201-11300batch: iter_time=2.340, loss_ctc=224.236, loss_att=187.780, acc=0.269, loss=198.690, grad_norm=4.378, loss_scale=1.000, learning_rate=1.063e-04, step_time=0.437 [cnode7-008:0/16] 2024-11-28 21:13:05,144 (deepspeed_trainer:226) INFO: 3epoch:train:11301-11400batch: iter_time=1.050e-04, loss_ctc=178.757, loss_att=152.340, acc=0.285, loss=160.268, grad_norm=4.182, loss_scale=1.000, learning_rate=1.068e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-28 21:13:43,157 (deepspeed_trainer:226) INFO: 3epoch:train:11401-11500batch: iter_time=1.086e-04, loss_ctc=201.796, loss_att=171.959, acc=0.268, loss=180.886, grad_norm=4.098, loss_scale=1.000, learning_rate=1.073e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-28 21:14:22,226 (deepspeed_trainer:226) INFO: 3epoch:train:11501-11600batch: iter_time=1.128e-04, loss_ctc=216.531, loss_att=186.310, acc=0.271, loss=195.330, grad_norm=4.136, loss_scale=1.000, learning_rate=1.078e-04, step_time=0.390 [cnode7-008:0/16] 2024-11-28 21:15:00,301 (deepspeed_trainer:226) INFO: 3epoch:train:11601-11700batch: iter_time=1.083e-04, loss_ctc=205.165, loss_att=174.376, acc=0.271, loss=183.624, grad_norm=4.398, loss_scale=1.000, learning_rate=1.083e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 21:15:38,298 (deepspeed_trainer:226) INFO: 3epoch:train:11701-11800batch: iter_time=1.093e-04, loss_ctc=183.838, loss_att=153.464, acc=0.276, loss=162.525, grad_norm=3.619, loss_scale=1.000, learning_rate=1.088e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-28 21:16:15,604 (deepspeed_trainer:226) INFO: 3epoch:train:11801-11900batch: iter_time=1.089e-04, loss_ctc=218.489, loss_att=181.662, acc=0.267, loss=192.685, grad_norm=4.684, loss_scale=1.000, learning_rate=1.093e-04, step_time=0.373 [2024-11-28 21:16:52,265] [INFO] [logging.py:129:log_dist] [Rank 0] step=42000, skipped=0, lr=[np.float64(0.000110005)], mom=[[0.9, 0.98]] [2024-11-28 21:16:52,266] [INFO] [timer.py:264:stop] epoch=0/micro_step=42000/global_step=42000, RunningAvgSamplesPerSec=42.7844773110367, CurrSamplesPerSec=44.55235461869232, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:16:52,267 (deepspeed_trainer:226) INFO: 3epoch:train:11901-12000batch: iter_time=1.077e-04, loss_ctc=209.175, loss_att=176.696, acc=0.273, loss=186.410, grad_norm=4.500, loss_scale=1.000, learning_rate=1.098e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 21:17:28,898 (deepspeed_trainer:226) INFO: 3epoch:train:12001-12100batch: iter_time=1.106e-04, loss_ctc=200.085, loss_att=172.650, acc=0.270, loss=180.850, grad_norm=4.044, loss_scale=1.000, learning_rate=1.103e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 21:18:06,350 (deepspeed_trainer:226) INFO: 3epoch:train:12101-12200batch: iter_time=1.062e-04, loss_ctc=181.757, loss_att=164.523, acc=0.280, loss=169.720, grad_norm=3.661, loss_scale=1.000, learning_rate=1.108e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 21:18:43,248 (deepspeed_trainer:226) INFO: 3epoch:train:12201-12300batch: iter_time=1.172e-04, loss_ctc=198.587, loss_att=168.015, acc=0.273, loss=177.209, grad_norm=4.034, loss_scale=1.000, learning_rate=1.113e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 21:19:20,365 (deepspeed_trainer:226) INFO: 3epoch:train:12301-12400batch: iter_time=1.154e-04, loss_ctc=195.106, loss_att=165.509, acc=0.269, loss=174.384, grad_norm=3.898, loss_scale=1.000, learning_rate=1.118e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 21:19:57,679 (deepspeed_trainer:226) INFO: 3epoch:train:12401-12500batch: iter_time=1.162e-04, loss_ctc=207.775, loss_att=181.400, acc=0.271, loss=189.329, grad_norm=3.627, loss_scale=1.000, learning_rate=1.123e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 21:20:35,846 (deepspeed_trainer:226) INFO: 3epoch:train:12501-12600batch: iter_time=1.153e-04, loss_ctc=199.979, loss_att=172.040, acc=0.275, loss=180.284, grad_norm=3.615, loss_scale=1.000, learning_rate=1.128e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 21:21:13,076 (deepspeed_trainer:226) INFO: 3epoch:train:12601-12700batch: iter_time=1.128e-04, loss_ctc=222.680, loss_att=179.740, acc=0.278, loss=192.623, grad_norm=4.468, loss_scale=1.000, learning_rate=1.133e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 21:21:50,817 (deepspeed_trainer:226) INFO: 3epoch:train:12701-12800batch: iter_time=1.171e-04, loss_ctc=214.684, loss_att=188.056, acc=0.262, loss=196.062, grad_norm=4.350, loss_scale=1.000, learning_rate=1.138e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 21:22:28,862 (deepspeed_trainer:226) INFO: 3epoch:train:12801-12900batch: iter_time=1.141e-04, loss_ctc=192.398, loss_att=165.933, acc=0.280, loss=173.860, grad_norm=3.838, loss_scale=1.000, learning_rate=1.143e-04, step_time=0.380 [2024-11-28 21:23:06,917] [INFO] [logging.py:129:log_dist] [Rank 0] step=43000, skipped=0, lr=[np.float64(0.000115005)], mom=[[0.9, 0.98]] [2024-11-28 21:23:06,918] [INFO] [timer.py:264:stop] epoch=0/micro_step=43000/global_step=43000, RunningAvgSamplesPerSec=42.80197667144502, CurrSamplesPerSec=43.62011923714277, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:23:06,919 (deepspeed_trainer:226) INFO: 3epoch:train:12901-13000batch: iter_time=1.100e-04, loss_ctc=195.350, loss_att=161.607, acc=0.278, loss=171.718, grad_norm=4.114, loss_scale=1.000, learning_rate=1.148e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-28 21:23:46,474 (deepspeed_trainer:226) INFO: 3epoch:train:13001-13100batch: iter_time=1.110e-04, loss_ctc=206.459, loss_att=182.049, acc=0.271, loss=189.434, grad_norm=3.906, loss_scale=1.000, learning_rate=1.153e-04, step_time=0.391 [cnode7-008:0/16] 2024-11-28 21:23:59,088 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-28 21:24:25,722 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 21:24:31,264 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 21:24:31,264 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-28 21:24:31,267 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 21:24:53,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:53,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:59,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:59,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:54,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:54,236] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:00,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:55,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:00,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:00,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:55,334] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:01,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:55,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:24:55,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:01,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:01,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:48,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:42,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:48,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:43,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:48,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:43,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:43,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:49,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:43,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:44,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:49,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:44,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:44,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:50,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:52,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:25:57,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:36,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:31,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:37,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:32,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:32,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:38,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:38,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:32,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:38,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:33,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:39,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:33,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:33,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:33,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:42,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:26:48,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:20,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:26,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:26,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:20,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:20,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:21,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:27,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:27,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:21,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:21,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:28,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:28,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:22,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:22,989] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:31,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:27:38,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 21:28:30,737 (deepspeed_trainer:226) INFO: 3epoch:train:13101-13200batch: iter_time=2.368, loss_ctc=210.786, loss_att=189.362, acc=0.275, loss=195.782, grad_norm=4.014, loss_scale=1.000, learning_rate=1.158e-04, step_time=0.478 [cnode7-008:0/16] 2024-11-28 21:29:07,990 (deepspeed_trainer:226) INFO: 3epoch:train:13201-13300batch: iter_time=1.049e-04, loss_ctc=179.939, loss_att=157.091, acc=0.274, loss=163.990, grad_norm=3.849, loss_scale=1.000, learning_rate=1.163e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 21:29:44,913 (deepspeed_trainer:226) INFO: 3epoch:train:13301-13400batch: iter_time=1.027e-04, loss_ctc=200.906, loss_att=175.665, acc=0.270, loss=183.301, grad_norm=3.718, loss_scale=1.000, learning_rate=1.168e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 21:30:22,149 (deepspeed_trainer:226) INFO: 3epoch:train:13401-13500batch: iter_time=1.086e-04, loss_ctc=209.102, loss_att=194.869, acc=0.272, loss=199.116, grad_norm=3.957, loss_scale=1.000, learning_rate=1.173e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 21:30:59,886 (deepspeed_trainer:226) INFO: 3epoch:train:13501-13600batch: iter_time=1.067e-04, loss_ctc=198.900, loss_att=172.187, acc=0.270, loss=180.284, grad_norm=3.944, loss_scale=1.000, learning_rate=1.178e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 21:31:37,048 (deepspeed_trainer:226) INFO: 3epoch:train:13601-13700batch: iter_time=1.088e-04, loss_ctc=172.451, loss_att=153.745, acc=0.280, loss=159.371, grad_norm=3.775, loss_scale=1.000, learning_rate=1.183e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 21:32:14,412 (deepspeed_trainer:226) INFO: 3epoch:train:13701-13800batch: iter_time=1.085e-04, loss_ctc=220.241, loss_att=184.667, acc=0.264, loss=195.445, grad_norm=5.389, loss_scale=1.000, learning_rate=1.188e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 21:32:51,636 (deepspeed_trainer:226) INFO: 3epoch:train:13801-13900batch: iter_time=1.094e-04, loss_ctc=212.049, loss_att=186.460, acc=0.269, loss=194.045, grad_norm=4.439, loss_scale=1.000, learning_rate=1.193e-04, step_time=0.372 [2024-11-28 21:33:29,779] [INFO] [logging.py:129:log_dist] [Rank 0] step=44000, skipped=0, lr=[np.float64(0.00012000500000000001)], mom=[[0.9, 0.98]] [2024-11-28 21:33:29,780] [INFO] [timer.py:264:stop] epoch=0/micro_step=44000/global_step=44000, RunningAvgSamplesPerSec=42.78916184870181, CurrSamplesPerSec=41.96850453953724, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:33:29,781 (deepspeed_trainer:226) INFO: 3epoch:train:13901-14000batch: iter_time=1.084e-04, loss_ctc=188.619, loss_att=169.593, acc=0.275, loss=175.299, grad_norm=4.246, loss_scale=1.000, learning_rate=1.198e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 21:34:08,682 (deepspeed_trainer:226) INFO: 3epoch:train:14001-14100batch: iter_time=1.090e-04, loss_ctc=178.720, loss_att=163.849, acc=0.283, loss=168.277, grad_norm=3.890, loss_scale=1.000, learning_rate=1.203e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-28 21:34:46,171 (deepspeed_trainer:226) INFO: 3epoch:train:14101-14200batch: iter_time=1.111e-04, loss_ctc=197.315, loss_att=175.556, acc=0.267, loss=182.067, grad_norm=3.953, loss_scale=1.000, learning_rate=1.208e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 21:35:24,240 (deepspeed_trainer:226) INFO: 3epoch:train:14201-14300batch: iter_time=1.103e-04, loss_ctc=186.025, loss_att=163.305, acc=0.277, loss=170.121, grad_norm=4.079, loss_scale=1.000, learning_rate=1.213e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 21:36:01,511 (deepspeed_trainer:226) INFO: 3epoch:train:14301-14400batch: iter_time=1.087e-04, loss_ctc=208.805, loss_att=188.735, acc=0.269, loss=194.610, grad_norm=3.485, loss_scale=1.000, learning_rate=1.218e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 21:36:38,357 (deepspeed_trainer:226) INFO: 3epoch:train:14401-14500batch: iter_time=1.080e-04, loss_ctc=192.971, loss_att=167.789, acc=0.285, loss=175.254, grad_norm=3.735, loss_scale=1.000, learning_rate=1.223e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 21:37:15,386 (deepspeed_trainer:226) INFO: 3epoch:train:14501-14600batch: iter_time=1.079e-04, loss_ctc=214.928, loss_att=185.275, acc=0.269, loss=194.132, grad_norm=4.394, loss_scale=1.000, learning_rate=1.228e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 21:37:52,520 (deepspeed_trainer:226) INFO: 3epoch:train:14601-14700batch: iter_time=1.081e-04, loss_ctc=208.225, loss_att=191.393, acc=0.266, loss=196.565, grad_norm=4.195, loss_scale=1.000, learning_rate=1.233e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 21:38:30,100 (deepspeed_trainer:226) INFO: 3epoch:train:14701-14800batch: iter_time=1.062e-04, loss_ctc=182.276, loss_att=164.064, acc=0.286, loss=169.501, grad_norm=3.179, loss_scale=1.000, learning_rate=1.238e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 21:39:07,221 (deepspeed_trainer:226) INFO: 3epoch:train:14801-14900batch: iter_time=1.038e-04, loss_ctc=201.278, loss_att=177.796, acc=0.273, loss=184.870, grad_norm=3.698, loss_scale=1.000, learning_rate=1.243e-04, step_time=0.372 [2024-11-28 21:39:44,407] [INFO] [logging.py:129:log_dist] [Rank 0] step=45000, skipped=0, lr=[np.float64(0.000125005)], mom=[[0.9, 0.98]] [2024-11-28 21:39:44,407] [INFO] [timer.py:264:stop] epoch=0/micro_step=45000/global_step=45000, RunningAvgSamplesPerSec=42.805459976857385, CurrSamplesPerSec=42.1042150852373, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:39:44,409 (deepspeed_trainer:226) INFO: 3epoch:train:14901-15000batch: iter_time=1.066e-04, loss_ctc=202.985, loss_att=185.979, acc=0.275, loss=191.083, grad_norm=3.993, loss_scale=1.000, learning_rate=1.248e-04, step_time=0.371 [2024-11-28 21:39:59,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:59,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:53,627] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:00,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:00,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:00,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:39:54,885] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:00,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:01,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:01,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:09,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:15,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:09,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:15,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:15,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:16,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:10,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:11,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:17,085] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:11,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:11,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:17,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:11,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:11,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:17,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:17,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:31,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:25,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:25,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:31,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:31,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:26,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:32,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:26,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:32,450] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:32,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:33,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:27,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:27,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:33,329] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:27,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:27,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:40,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:46,524] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:41,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:47,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:41,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:42,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:48,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:42,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:48,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:42,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:48,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:49,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:43,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:49,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:49,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:40:43,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:01,755] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 3 is about to be saved! [2024-11-28 21:41:01,786] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/mp_rank_00_model_states.pt [2024-11-28 21:41:01,786] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/mp_rank_00_model_states.pt... [2024-11-28 21:41:03,776] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/mp_rank_00_model_states.pt. [2024-11-28 21:41:03,934] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-28 21:40:57,994] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-28 21:40:57,995] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,939] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,939] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,941] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,941] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,941] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,941] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,001] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,001] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,002] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-28 21:41:03,943] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,002] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,002] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-28 21:40:58,003] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-28 21:41:04,535] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,535] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,535] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,548] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,549] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,549] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,673] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,673] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,673] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,673] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,673] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,673] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,688] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,688] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,689] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,694] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,694] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,694] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,757] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,757] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,757] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,758] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,758] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,759] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,709] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,715] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,715] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,715] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,726] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,726] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,789] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,789] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,789] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,795] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,796] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,796] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:40:58,802] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-28 21:40:58,802] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-28 21:40:58,802] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,794] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,794] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,794] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,799] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,799] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,799] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [2024-11-28 21:41:04,817] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-28 21:41:04,818] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_3/3/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-28 21:41:04,818] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 3 is ready now! [cnode7-008:0/16] 2024-11-28 21:41:04,875 (deepspeed_trainer:158) INFO: 3epoch results: [train] iter_time=0.124, loss_ctc=214.778, loss_att=180.495, acc=0.264, loss=190.781, grad_norm=5.314, loss_scale=1.000, learning_rate=8.751e-05, step_time=0.379, time=2 hours, 5 minutes and 56.21 seconds, total_count=45003, gpu_max_cached_mem_GB=27.396, [valid] loss_ctc=41.500, cer_ctc=0.979, loss_att=55.000, acc=0.375, cer=0.611, wer=1.000, loss=51.000, time=1 minute and 8.37 seconds, total_count=3, gpu_max_cached_mem_GB=27.396 [cnode7-008:0/16] 2024-11-28 21:41:05,877 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-28 21:41:32,320 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 21:41:37,723 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 21:41:37,723 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-28 21:41:37,725 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 21:42:00,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:00,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:54,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:55,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:01,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:55,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:56,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:02,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:56,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:56,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:56,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:02,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:41:56,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:02,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:03,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:03,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:46,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:48,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:42,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:49,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:43,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:43,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:44,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:50,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:44,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:44,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:44,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:50,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:45,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:51,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:51,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:42:53,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:33,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:35,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:36,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:30,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:36,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:31,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:31,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:31,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:31,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:32,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:38,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:38,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:33,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:39,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:33,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:43:42,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:19,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:22,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:22,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:17,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:18,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:24,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:18,759] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:25,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:19,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:25,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:25,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:19,968] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:20,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:22,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:22,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:44:30,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 21:45:33,609 (deepspeed_trainer:226) INFO: 4epoch:train:1-100batch: iter_time=2.204, loss_ctc=188.036, loss_att=159.713, acc=0.283, loss=168.285, grad_norm=3.957, loss_scale=1.000, learning_rate=1.253e-04, step_time=0.474 [cnode7-008:0/16] 2024-11-28 21:46:13,176 (deepspeed_trainer:226) INFO: 4epoch:train:101-200batch: iter_time=1.086e-04, loss_ctc=215.785, loss_att=196.632, acc=0.265, loss=202.415, grad_norm=4.495, loss_scale=1.000, learning_rate=1.258e-04, step_time=0.396 [cnode7-008:0/16] 2024-11-28 21:46:50,752 (deepspeed_trainer:226) INFO: 4epoch:train:201-300batch: iter_time=1.094e-04, loss_ctc=199.887, loss_att=173.567, acc=0.274, loss=181.445, grad_norm=4.258, loss_scale=1.000, learning_rate=1.263e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 21:47:28,970 (deepspeed_trainer:226) INFO: 4epoch:train:301-400batch: iter_time=1.078e-04, loss_ctc=203.821, loss_att=183.208, acc=0.269, loss=189.285, grad_norm=3.703, loss_scale=1.000, learning_rate=1.268e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-28 21:48:05,718 (deepspeed_trainer:226) INFO: 4epoch:train:401-500batch: iter_time=1.070e-04, loss_ctc=186.702, loss_att=161.243, acc=0.278, loss=168.946, grad_norm=3.458, loss_scale=1.000, learning_rate=1.273e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 21:48:42,609 (deepspeed_trainer:226) INFO: 4epoch:train:501-600batch: iter_time=1.091e-04, loss_ctc=175.268, loss_att=155.143, acc=0.288, loss=161.143, grad_norm=3.292, loss_scale=1.000, learning_rate=1.278e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 21:49:19,752 (deepspeed_trainer:226) INFO: 4epoch:train:601-700batch: iter_time=1.115e-04, loss_ctc=183.025, loss_att=160.421, acc=0.291, loss=167.171, grad_norm=3.723, loss_scale=1.000, learning_rate=1.283e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 21:49:56,624 (deepspeed_trainer:226) INFO: 4epoch:train:701-800batch: iter_time=1.075e-04, loss_ctc=188.327, loss_att=165.666, acc=0.284, loss=172.534, grad_norm=4.067, loss_scale=1.000, learning_rate=1.288e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 21:50:33,936 (deepspeed_trainer:226) INFO: 4epoch:train:801-900batch: iter_time=1.075e-04, loss_ctc=204.428, loss_att=180.343, acc=0.264, loss=187.523, grad_norm=4.825, loss_scale=1.000, learning_rate=1.293e-04, step_time=0.373 [2024-11-28 21:51:11,266] [INFO] [logging.py:129:log_dist] [Rank 0] step=46000, skipped=0, lr=[np.float64(0.000130005)], mom=[[0.9, 0.98]] [2024-11-28 21:51:11,266] [INFO] [timer.py:264:stop] epoch=0/micro_step=46000/global_step=46000, RunningAvgSamplesPerSec=42.79550405709781, CurrSamplesPerSec=43.87530301663585, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 21:51:11,268 (deepspeed_trainer:226) INFO: 4epoch:train:901-1000batch: iter_time=1.077e-04, loss_ctc=183.488, loss_att=166.234, acc=0.275, loss=171.419, grad_norm=4.151, loss_scale=1.000, learning_rate=1.298e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 21:51:48,102 (deepspeed_trainer:226) INFO: 4epoch:train:1001-1100batch: iter_time=1.089e-04, loss_ctc=191.811, loss_att=171.600, acc=0.283, loss=177.635, grad_norm=3.622, loss_scale=1.000, learning_rate=1.303e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 21:52:25,957 (deepspeed_trainer:226) INFO: 4epoch:train:1101-1200batch: iter_time=1.082e-04, loss_ctc=199.208, loss_att=185.686, acc=0.275, loss=189.694, grad_norm=3.851, loss_scale=1.000, learning_rate=1.308e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 21:53:04,271 (deepspeed_trainer:226) INFO: 4epoch:train:1201-1300batch: iter_time=1.092e-04, loss_ctc=168.825, loss_att=155.980, acc=0.286, loss=159.812, grad_norm=3.687, loss_scale=1.000, learning_rate=1.313e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-28 21:53:43,242 (deepspeed_trainer:226) INFO: 4epoch:train:1301-1400batch: iter_time=1.078e-04, loss_ctc=180.873, loss_att=163.085, acc=0.283, loss=168.393, grad_norm=4.294, loss_scale=1.000, learning_rate=1.318e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-28 21:54:20,898 (deepspeed_trainer:226) INFO: 4epoch:train:1401-1500batch: iter_time=1.083e-04, loss_ctc=200.564, loss_att=186.784, acc=0.272, loss=190.824, grad_norm=3.867, loss_scale=1.000, learning_rate=1.323e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 21:54:59,440 (deepspeed_trainer:226) INFO: 4epoch:train:1501-1600batch: iter_time=1.099e-04, loss_ctc=174.286, loss_att=160.879, acc=0.288, loss=164.925, grad_norm=3.301, loss_scale=1.000, learning_rate=1.328e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-28 21:55:36,747 (deepspeed_trainer:226) INFO: 4epoch:train:1601-1700batch: iter_time=1.116e-04, loss_ctc=173.604, loss_att=161.031, acc=0.287, loss=164.821, grad_norm=3.825, loss_scale=1.000, learning_rate=1.333e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 21:56:13,799 (deepspeed_trainer:226) INFO: 4epoch:train:1701-1800batch: iter_time=1.083e-04, loss_ctc=191.379, loss_att=172.243, acc=0.277, loss=177.980, grad_norm=3.980, loss_scale=1.000, learning_rate=1.338e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 21:56:45,634 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-28 21:57:12,203 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 21:57:18,633 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 21:57:18,634 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-28 21:57:18,636 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 21:57:43,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:43,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:43,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:43,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:44,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:44,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:44,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:44,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:51,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,000] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:57:52,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:35,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:35,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:36,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:36,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:36,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:36,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:36,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:38,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:45,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:46,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:46,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:46,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:47,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:47,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:47,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:58:47,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:24,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:24,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:25,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:26,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:26,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:26,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:26,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:34,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:28,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:34,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:35,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:35,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:35,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:36,280] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:36,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 21:59:36,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:11,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:12,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:13,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:14,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:14,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:15,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:15,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:22,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:22,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:23,700] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:24,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:18,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:24,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:24,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:25,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:00:25,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 22:01:01,998 (deepspeed_trainer:226) INFO: 4epoch:train:1801-1900batch: iter_time=2.478, loss_ctc=192.706, loss_att=179.456, acc=0.283, loss=183.380, grad_norm=4.399, loss_scale=1.000, learning_rate=1.343e-04, step_time=0.404 [2024-11-28 22:01:39,767] [INFO] [logging.py:129:log_dist] [Rank 0] step=47000, skipped=0, lr=[np.float64(0.000135005)], mom=[[0.9, 0.98]] [2024-11-28 22:01:39,770] [INFO] [timer.py:264:stop] epoch=0/micro_step=47000/global_step=47000, RunningAvgSamplesPerSec=42.79630653263389, CurrSamplesPerSec=45.32673167335203, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:01:39,771 (deepspeed_trainer:226) INFO: 4epoch:train:1901-2000batch: iter_time=1.067e-04, loss_ctc=188.817, loss_att=178.666, acc=0.276, loss=181.661, grad_norm=3.921, loss_scale=1.000, learning_rate=1.348e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 22:02:17,485 (deepspeed_trainer:226) INFO: 4epoch:train:2001-2100batch: iter_time=1.099e-04, loss_ctc=201.450, loss_att=187.415, acc=0.271, loss=191.625, grad_norm=4.311, loss_scale=1.000, learning_rate=1.353e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 22:02:55,285 (deepspeed_trainer:226) INFO: 4epoch:train:2101-2200batch: iter_time=1.081e-04, loss_ctc=187.079, loss_att=177.246, acc=0.276, loss=180.240, grad_norm=4.043, loss_scale=1.000, learning_rate=1.358e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 22:03:34,273 (deepspeed_trainer:226) INFO: 4epoch:train:2201-2300batch: iter_time=1.090e-04, loss_ctc=188.824, loss_att=183.321, acc=0.271, loss=185.001, grad_norm=3.572, loss_scale=1.000, learning_rate=1.363e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-28 22:04:11,243 (deepspeed_trainer:226) INFO: 4epoch:train:2301-2400batch: iter_time=1.081e-04, loss_ctc=174.042, loss_att=161.845, acc=0.287, loss=165.505, grad_norm=3.756, loss_scale=1.000, learning_rate=1.368e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 22:04:49,144 (deepspeed_trainer:226) INFO: 4epoch:train:2401-2500batch: iter_time=1.074e-04, loss_ctc=158.991, loss_att=151.785, acc=0.291, loss=153.956, grad_norm=3.611, loss_scale=1.000, learning_rate=1.373e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 22:05:26,213 (deepspeed_trainer:226) INFO: 4epoch:train:2501-2600batch: iter_time=1.080e-04, loss_ctc=184.327, loss_att=173.654, acc=0.291, loss=176.792, grad_norm=3.766, loss_scale=1.000, learning_rate=1.378e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 22:06:02,867 (deepspeed_trainer:226) INFO: 4epoch:train:2601-2700batch: iter_time=1.077e-04, loss_ctc=165.606, loss_att=153.976, acc=0.289, loss=157.463, grad_norm=3.623, loss_scale=1.000, learning_rate=1.383e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 22:06:40,061 (deepspeed_trainer:226) INFO: 4epoch:train:2701-2800batch: iter_time=1.098e-04, loss_ctc=190.368, loss_att=180.759, acc=0.269, loss=183.639, grad_norm=4.043, loss_scale=1.000, learning_rate=1.388e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:07:16,883 (deepspeed_trainer:226) INFO: 4epoch:train:2801-2900batch: iter_time=1.073e-04, loss_ctc=166.523, loss_att=163.106, acc=0.291, loss=164.116, grad_norm=4.005, loss_scale=1.000, learning_rate=1.393e-04, step_time=0.368 [2024-11-28 22:07:53,939] [INFO] [logging.py:129:log_dist] [Rank 0] step=48000, skipped=0, lr=[np.float64(0.000140005)], mom=[[0.9, 0.98]] [2024-11-28 22:07:53,939] [INFO] [timer.py:264:stop] epoch=0/micro_step=48000/global_step=48000, RunningAvgSamplesPerSec=42.81301993202775, CurrSamplesPerSec=39.58512262303971, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:07:53,941 (deepspeed_trainer:226) INFO: 4epoch:train:2901-3000batch: iter_time=1.072e-04, loss_ctc=185.452, loss_att=183.396, acc=0.282, loss=183.910, grad_norm=3.813, loss_scale=1.000, learning_rate=1.398e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 22:08:31,802 (deepspeed_trainer:226) INFO: 4epoch:train:3001-3100batch: iter_time=1.043e-04, loss_ctc=184.110, loss_att=186.279, acc=0.282, loss=185.609, grad_norm=3.890, loss_scale=1.000, learning_rate=1.403e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-28 22:09:08,882 (deepspeed_trainer:226) INFO: 4epoch:train:3101-3200batch: iter_time=1.056e-04, loss_ctc=154.475, loss_att=150.178, acc=0.292, loss=151.501, grad_norm=4.115, loss_scale=1.000, learning_rate=1.408e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 22:09:46,389 (deepspeed_trainer:226) INFO: 4epoch:train:3201-3300batch: iter_time=1.064e-04, loss_ctc=176.150, loss_att=171.690, acc=0.282, loss=172.990, grad_norm=4.348, loss_scale=1.000, learning_rate=1.413e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 22:10:24,295 (deepspeed_trainer:226) INFO: 4epoch:train:3301-3400batch: iter_time=1.083e-04, loss_ctc=176.220, loss_att=175.548, acc=0.287, loss=175.815, grad_norm=3.884, loss_scale=1.000, learning_rate=1.418e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-28 22:11:02,476 (deepspeed_trainer:226) INFO: 4epoch:train:3401-3500batch: iter_time=1.086e-04, loss_ctc=166.946, loss_att=161.120, acc=0.295, loss=162.827, grad_norm=3.489, loss_scale=1.000, learning_rate=1.423e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 22:11:41,332 (deepspeed_trainer:226) INFO: 4epoch:train:3501-3600batch: iter_time=1.077e-04, loss_ctc=155.095, loss_att=156.031, acc=0.297, loss=155.851, grad_norm=3.494, loss_scale=1.000, learning_rate=1.428e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-28 22:12:18,750 (deepspeed_trainer:226) INFO: 4epoch:train:3601-3700batch: iter_time=1.074e-04, loss_ctc=188.643, loss_att=184.794, acc=0.285, loss=185.974, grad_norm=3.896, loss_scale=1.000, learning_rate=1.433e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 22:12:41,494 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-28 22:13:08,606 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 22:13:14,221 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 22:13:14,221 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-28 22:13:14,224 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 22:13:42,359] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:42,594] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:42,710] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:37,714] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:43,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:43,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:43,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:43,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:13:43,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:24,791] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:30,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:30,625] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:25,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:30,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:30,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:26,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:31,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:26,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:26,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:32,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:32,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:27,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:27,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:27,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:14:33,113] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:12,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:18,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:13,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:19,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:19,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:19,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:14,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:19,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:14,813] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:15,150] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:20,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:20,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:15,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:21,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:15,945] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:17,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:04,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:15:59,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:05,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:05,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:00,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:06,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:01,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:02,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:07,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:07,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:02,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:03,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:09,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:04,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:09,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:16:07,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 22:16:57,303 (deepspeed_trainer:226) INFO: 4epoch:train:3701-3800batch: iter_time=2.338, loss_ctc=191.076, loss_att=185.421, acc=0.279, loss=187.060, grad_norm=4.601, loss_scale=1.000, learning_rate=1.438e-04, step_time=0.447 [cnode7-008:0/16] 2024-11-28 22:17:34,014 (deepspeed_trainer:226) INFO: 4epoch:train:3801-3900batch: iter_time=1.047e-04, loss_ctc=162.896, loss_att=159.095, acc=0.296, loss=160.264, grad_norm=3.665, loss_scale=1.000, learning_rate=1.443e-04, step_time=0.367 [2024-11-28 22:18:11,565] [INFO] [logging.py:129:log_dist] [Rank 0] step=49000, skipped=0, lr=[np.float64(0.000145005)], mom=[[0.9, 0.98]] [2024-11-28 22:18:11,566] [INFO] [timer.py:264:stop] epoch=0/micro_step=49000/global_step=49000, RunningAvgSamplesPerSec=42.80696529391457, CurrSamplesPerSec=46.542284748939174, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:18:11,567 (deepspeed_trainer:226) INFO: 4epoch:train:3901-4000batch: iter_time=1.056e-04, loss_ctc=189.716, loss_att=183.273, acc=0.283, loss=185.238, grad_norm=4.524, loss_scale=1.000, learning_rate=1.448e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 22:18:48,835 (deepspeed_trainer:226) INFO: 4epoch:train:4001-4100batch: iter_time=1.037e-04, loss_ctc=172.665, loss_att=172.537, acc=0.288, loss=172.639, grad_norm=3.686, loss_scale=1.000, learning_rate=1.453e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:19:26,342 (deepspeed_trainer:226) INFO: 4epoch:train:4101-4200batch: iter_time=1.093e-04, loss_ctc=169.875, loss_att=161.364, acc=0.295, loss=163.919, grad_norm=3.773, loss_scale=1.000, learning_rate=1.458e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 22:20:04,100 (deepspeed_trainer:226) INFO: 4epoch:train:4201-4300batch: iter_time=1.058e-04, loss_ctc=163.710, loss_att=155.840, acc=0.298, loss=158.204, grad_norm=3.924, loss_scale=1.000, learning_rate=1.463e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 22:20:41,333 (deepspeed_trainer:226) INFO: 4epoch:train:4301-4400batch: iter_time=1.068e-04, loss_ctc=148.755, loss_att=145.665, acc=0.303, loss=146.599, grad_norm=4.257, loss_scale=1.000, learning_rate=1.468e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:21:20,317 (deepspeed_trainer:226) INFO: 4epoch:train:4401-4500batch: iter_time=1.082e-04, loss_ctc=175.959, loss_att=168.206, acc=0.304, loss=170.464, grad_norm=3.700, loss_scale=1.000, learning_rate=1.473e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-28 22:21:57,914 (deepspeed_trainer:226) INFO: 4epoch:train:4501-4600batch: iter_time=1.060e-04, loss_ctc=163.244, loss_att=153.055, acc=0.294, loss=156.120, grad_norm=3.795, loss_scale=1.000, learning_rate=1.478e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 22:22:36,321 (deepspeed_trainer:226) INFO: 4epoch:train:4601-4700batch: iter_time=1.080e-04, loss_ctc=167.420, loss_att=162.699, acc=0.294, loss=164.114, grad_norm=4.063, loss_scale=1.000, learning_rate=1.483e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-28 22:23:13,492 (deepspeed_trainer:226) INFO: 4epoch:train:4701-4800batch: iter_time=1.070e-04, loss_ctc=166.321, loss_att=163.110, acc=0.300, loss=164.123, grad_norm=4.334, loss_scale=1.000, learning_rate=1.488e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 22:23:50,711 (deepspeed_trainer:226) INFO: 4epoch:train:4801-4900batch: iter_time=1.076e-04, loss_ctc=171.559, loss_att=173.606, acc=0.298, loss=173.060, grad_norm=4.465, loss_scale=1.000, learning_rate=1.493e-04, step_time=0.372 [2024-11-28 22:24:27,911] [INFO] [logging.py:129:log_dist] [Rank 0] step=50000, skipped=0, lr=[np.float64(0.000150005)], mom=[[0.9, 0.98]] [2024-11-28 22:24:27,912] [INFO] [timer.py:264:stop] epoch=0/micro_step=50000/global_step=50000, RunningAvgSamplesPerSec=42.81772285514602, CurrSamplesPerSec=46.398573369607206, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:24:27,913 (deepspeed_trainer:226) INFO: 4epoch:train:4901-5000batch: iter_time=1.066e-04, loss_ctc=163.829, loss_att=168.645, acc=0.297, loss=167.268, grad_norm=4.340, loss_scale=1.000, learning_rate=1.498e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:25:05,014 (deepspeed_trainer:226) INFO: 4epoch:train:5001-5100batch: iter_time=1.070e-04, loss_ctc=151.738, loss_att=145.482, acc=0.307, loss=147.346, grad_norm=3.811, loss_scale=1.000, learning_rate=1.503e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 22:25:41,937 (deepspeed_trainer:226) INFO: 4epoch:train:5101-5200batch: iter_time=1.074e-04, loss_ctc=167.250, loss_att=167.905, acc=0.299, loss=167.625, grad_norm=3.832, loss_scale=1.000, learning_rate=1.508e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 22:26:18,603 (deepspeed_trainer:226) INFO: 4epoch:train:5201-5300batch: iter_time=1.099e-04, loss_ctc=170.011, loss_att=170.639, acc=0.300, loss=170.433, grad_norm=3.644, loss_scale=1.000, learning_rate=1.513e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 22:26:55,980 (deepspeed_trainer:226) INFO: 4epoch:train:5301-5400batch: iter_time=1.062e-04, loss_ctc=144.861, loss_att=142.583, acc=0.316, loss=143.322, grad_norm=3.413, loss_scale=1.000, learning_rate=1.518e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 22:27:33,395 (deepspeed_trainer:226) INFO: 4epoch:train:5401-5500batch: iter_time=1.040e-04, loss_ctc=144.980, loss_att=147.904, acc=0.310, loss=147.002, grad_norm=3.113, loss_scale=1.000, learning_rate=1.523e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 22:28:10,393 (deepspeed_trainer:226) INFO: 4epoch:train:5501-5600batch: iter_time=1.058e-04, loss_ctc=186.213, loss_att=183.773, acc=0.295, loss=184.575, grad_norm=4.142, loss_scale=1.000, learning_rate=1.528e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 22:28:23,032 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-28 22:28:50,184 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 22:28:56,078 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 22:28:56,078 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-28 22:28:56,080 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 22:29:22,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:17,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:23,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:17,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:23,756] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:18,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:24,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:18,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:24,550] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:24,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:19,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:19,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:19,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:19,973] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:26,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:29:26,920] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:10,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:06,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:06,629] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:12,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:12,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:07,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:07,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:14,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:08,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:08,586] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:14,307] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:09,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:09,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:14,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:16,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:16,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:58,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:53,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:54,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:00,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:00,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:55,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:56,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:56,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:02,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:02,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:57,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:30:57,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:03,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:04,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:01,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:07,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:45,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:41,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:41,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:48,362] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:49,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:44,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:50,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:44,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:44,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:45,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:45,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:51,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:51,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:52,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:51,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:31:56,995] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 22:32:49,992 (deepspeed_trainer:226) INFO: 4epoch:train:5601-5700batch: iter_time=2.356, loss_ctc=164.350, loss_att=165.588, acc=0.300, loss=165.216, grad_norm=4.158, loss_scale=1.000, learning_rate=1.533e-04, step_time=0.439 [cnode7-008:0/16] 2024-11-28 22:33:27,231 (deepspeed_trainer:226) INFO: 4epoch:train:5701-5800batch: iter_time=1.047e-04, loss_ctc=158.505, loss_att=164.870, acc=0.299, loss=162.970, grad_norm=3.504, loss_scale=1.000, learning_rate=1.538e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:34:05,184 (deepspeed_trainer:226) INFO: 4epoch:train:5801-5900batch: iter_time=1.084e-04, loss_ctc=178.690, loss_att=182.090, acc=0.295, loss=181.269, grad_norm=4.417, loss_scale=1.000, learning_rate=1.543e-04, step_time=0.379 [2024-11-28 22:34:42,382] [INFO] [logging.py:129:log_dist] [Rank 0] step=51000, skipped=0, lr=[np.float64(0.000155005)], mom=[[0.9, 0.98]] [2024-11-28 22:34:42,383] [INFO] [timer.py:264:stop] epoch=0/micro_step=51000/global_step=51000, RunningAvgSamplesPerSec=42.82255034687653, CurrSamplesPerSec=43.79907524332005, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:34:42,384 (deepspeed_trainer:226) INFO: 4epoch:train:5901-6000batch: iter_time=1.090e-04, loss_ctc=160.454, loss_att=162.015, acc=0.309, loss=161.540, grad_norm=3.694, loss_scale=1.000, learning_rate=1.548e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:35:19,404 (deepspeed_trainer:226) INFO: 4epoch:train:6001-6100batch: iter_time=1.143e-04, loss_ctc=158.120, loss_att=156.809, acc=0.311, loss=157.106, grad_norm=3.564, loss_scale=1.000, learning_rate=1.553e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 22:35:56,356 (deepspeed_trainer:226) INFO: 4epoch:train:6101-6200batch: iter_time=1.129e-04, loss_ctc=143.727, loss_att=144.694, acc=0.316, loss=144.404, grad_norm=3.440, loss_scale=1.000, learning_rate=1.558e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 22:36:34,003 (deepspeed_trainer:226) INFO: 4epoch:train:6201-6300batch: iter_time=1.102e-04, loss_ctc=142.111, loss_att=146.321, acc=0.317, loss=145.054, grad_norm=3.511, loss_scale=1.000, learning_rate=1.563e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 22:37:12,406 (deepspeed_trainer:226) INFO: 4epoch:train:6301-6400batch: iter_time=1.126e-04, loss_ctc=162.550, loss_att=164.034, acc=0.323, loss=163.541, grad_norm=4.198, loss_scale=1.000, learning_rate=1.568e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-28 22:37:49,311 (deepspeed_trainer:226) INFO: 4epoch:train:6401-6500batch: iter_time=1.109e-04, loss_ctc=162.596, loss_att=162.310, acc=0.308, loss=162.347, grad_norm=4.288, loss_scale=1.000, learning_rate=1.573e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 22:38:27,133 (deepspeed_trainer:226) INFO: 4epoch:train:6501-6600batch: iter_time=1.069e-04, loss_ctc=150.623, loss_att=155.375, acc=0.312, loss=153.939, grad_norm=4.167, loss_scale=1.000, learning_rate=1.578e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 22:39:04,863 (deepspeed_trainer:226) INFO: 4epoch:train:6601-6700batch: iter_time=1.059e-04, loss_ctc=153.221, loss_att=162.976, acc=0.321, loss=160.080, grad_norm=3.590, loss_scale=1.000, learning_rate=1.583e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 22:39:44,223 (deepspeed_trainer:226) INFO: 4epoch:train:6701-6800batch: iter_time=1.063e-04, loss_ctc=159.619, loss_att=166.107, acc=0.313, loss=164.109, grad_norm=4.176, loss_scale=1.000, learning_rate=1.588e-04, step_time=0.394 [cnode7-008:0/16] 2024-11-28 22:40:22,437 (deepspeed_trainer:226) INFO: 4epoch:train:6801-6900batch: iter_time=1.080e-04, loss_ctc=152.463, loss_att=163.591, acc=0.317, loss=160.287, grad_norm=4.912, loss_scale=1.000, learning_rate=1.593e-04, step_time=0.382 [2024-11-28 22:41:00,085] [INFO] [logging.py:129:log_dist] [Rank 0] step=52000, skipped=0, lr=[np.float64(0.000160005)], mom=[[0.9, 0.98]] [2024-11-28 22:41:00,086] [INFO] [timer.py:264:stop] epoch=0/micro_step=52000/global_step=52000, RunningAvgSamplesPerSec=42.829781672484685, CurrSamplesPerSec=46.806622720247006, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:41:00,087 (deepspeed_trainer:226) INFO: 4epoch:train:6901-7000batch: iter_time=1.082e-04, loss_ctc=145.355, loss_att=145.343, acc=0.327, loss=145.421, grad_norm=5.031, loss_scale=1.000, learning_rate=1.598e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 22:41:37,799 (deepspeed_trainer:226) INFO: 4epoch:train:7001-7100batch: iter_time=1.093e-04, loss_ctc=156.955, loss_att=161.666, acc=0.318, loss=160.245, grad_norm=4.210, loss_scale=1.000, learning_rate=1.603e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 22:42:15,066 (deepspeed_trainer:226) INFO: 4epoch:train:7101-7200batch: iter_time=1.112e-04, loss_ctc=148.093, loss_att=153.991, acc=0.336, loss=152.257, grad_norm=3.782, loss_scale=1.000, learning_rate=1.608e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:42:52,058 (deepspeed_trainer:226) INFO: 4epoch:train:7201-7300batch: iter_time=1.110e-04, loss_ctc=138.759, loss_att=143.009, acc=0.332, loss=141.670, grad_norm=3.913, loss_scale=1.000, learning_rate=1.613e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 22:43:28,720 (deepspeed_trainer:226) INFO: 4epoch:train:7301-7400batch: iter_time=1.103e-04, loss_ctc=139.366, loss_att=143.553, acc=0.342, loss=142.300, grad_norm=4.660, loss_scale=1.000, learning_rate=1.618e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 22:44:05,920 (deepspeed_trainer:226) INFO: 4epoch:train:7401-7500batch: iter_time=1.101e-04, loss_ctc=171.157, loss_att=180.475, acc=0.313, loss=177.665, grad_norm=4.222, loss_scale=1.000, learning_rate=1.623e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:44:09,143 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-28 22:44:35,731 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 22:44:41,391 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 22:44:41,391 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-28 22:44:41,393 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 22:45:08,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:02,970] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:09,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:03,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:09,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:09,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:04,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:04,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:04,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:10,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:10,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:04,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:10,495] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:05,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:10,908] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:05,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:50,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:56,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:57,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:57,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:58,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:58,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:52,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:58,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:53,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:59,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:53,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:54,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:54,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:54,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:45:54,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:00,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:38,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:40,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:46,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:47,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:41,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:47,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:47,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:42,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:42,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:48,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:48,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:42,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:42,881] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:43,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:49,354] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:46:51,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:25,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:28,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:29,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:35,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:35,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:29,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:35,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:36,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:36,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:30,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:30,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:36,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:36,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:30,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:31,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 22:47:38,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 22:48:43,048 (deepspeed_trainer:226) INFO: 4epoch:train:7501-7600batch: iter_time=2.367, loss_ctc=146.407, loss_att=147.965, acc=0.322, loss=147.444, grad_norm=4.581, loss_scale=1.000, learning_rate=1.628e-04, step_time=0.404 [cnode7-008:0/16] 2024-11-28 22:49:20,561 (deepspeed_trainer:226) INFO: 4epoch:train:7601-7700batch: iter_time=1.034e-04, loss_ctc=163.445, loss_att=174.938, acc=0.319, loss=171.463, grad_norm=4.265, loss_scale=1.000, learning_rate=1.633e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 22:49:59,900 (deepspeed_trainer:226) INFO: 4epoch:train:7701-7800batch: iter_time=1.044e-04, loss_ctc=156.817, loss_att=152.684, acc=0.329, loss=153.907, grad_norm=4.085, loss_scale=1.000, learning_rate=1.638e-04, step_time=0.393 [cnode7-008:0/16] 2024-11-28 22:50:37,369 (deepspeed_trainer:226) INFO: 4epoch:train:7801-7900batch: iter_time=1.073e-04, loss_ctc=153.580, loss_att=155.210, acc=0.341, loss=154.715, grad_norm=3.843, loss_scale=1.000, learning_rate=1.643e-04, step_time=0.375 [2024-11-28 22:51:15,574] [INFO] [logging.py:129:log_dist] [Rank 0] step=53000, skipped=0, lr=[np.float64(0.000165005)], mom=[[0.9, 0.98]] [2024-11-28 22:51:15,574] [INFO] [timer.py:264:stop] epoch=0/micro_step=53000/global_step=53000, RunningAvgSamplesPerSec=42.83419863678454, CurrSamplesPerSec=43.595268120515605, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:51:15,576 (deepspeed_trainer:226) INFO: 4epoch:train:7901-8000batch: iter_time=1.103e-04, loss_ctc=144.034, loss_att=138.616, acc=0.345, loss=140.269, grad_norm=4.285, loss_scale=1.000, learning_rate=1.648e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 22:51:52,254 (deepspeed_trainer:226) INFO: 4epoch:train:8001-8100batch: iter_time=1.111e-04, loss_ctc=135.032, loss_att=137.114, acc=0.342, loss=136.468, grad_norm=4.423, loss_scale=1.000, learning_rate=1.653e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 22:52:29,144 (deepspeed_trainer:226) INFO: 4epoch:train:8101-8200batch: iter_time=1.098e-04, loss_ctc=138.838, loss_att=143.225, acc=0.348, loss=141.926, grad_norm=3.513, loss_scale=1.000, learning_rate=1.658e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 22:53:05,988 (deepspeed_trainer:226) INFO: 4epoch:train:8201-8300batch: iter_time=1.110e-04, loss_ctc=146.338, loss_att=145.664, acc=0.347, loss=145.870, grad_norm=3.745, loss_scale=1.000, learning_rate=1.663e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 22:53:43,338 (deepspeed_trainer:226) INFO: 4epoch:train:8301-8400batch: iter_time=1.123e-04, loss_ctc=157.925, loss_att=154.004, acc=0.334, loss=155.189, grad_norm=4.452, loss_scale=1.000, learning_rate=1.668e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 22:54:20,353 (deepspeed_trainer:226) INFO: 4epoch:train:8401-8500batch: iter_time=1.087e-04, loss_ctc=142.626, loss_att=149.488, acc=0.333, loss=147.425, grad_norm=4.246, loss_scale=1.000, learning_rate=1.673e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 22:54:57,184 (deepspeed_trainer:226) INFO: 4epoch:train:8501-8600batch: iter_time=1.081e-04, loss_ctc=142.906, loss_att=147.107, acc=0.360, loss=145.839, grad_norm=3.803, loss_scale=1.000, learning_rate=1.678e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 22:55:34,624 (deepspeed_trainer:226) INFO: 4epoch:train:8601-8700batch: iter_time=1.099e-04, loss_ctc=150.616, loss_att=161.161, acc=0.347, loss=157.961, grad_norm=4.122, loss_scale=1.000, learning_rate=1.683e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 22:56:12,073 (deepspeed_trainer:226) INFO: 4epoch:train:8701-8800batch: iter_time=1.089e-04, loss_ctc=128.809, loss_att=135.299, acc=0.349, loss=133.361, grad_norm=3.499, loss_scale=1.000, learning_rate=1.688e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 22:56:49,312 (deepspeed_trainer:226) INFO: 4epoch:train:8801-8900batch: iter_time=1.087e-04, loss_ctc=139.724, loss_att=140.819, acc=0.352, loss=140.490, grad_norm=4.211, loss_scale=1.000, learning_rate=1.693e-04, step_time=0.372 [2024-11-28 22:57:27,662] [INFO] [logging.py:129:log_dist] [Rank 0] step=54000, skipped=0, lr=[np.float64(0.000170005)], mom=[[0.9, 0.98]] [2024-11-28 22:57:27,663] [INFO] [timer.py:264:stop] epoch=0/micro_step=54000/global_step=54000, RunningAvgSamplesPerSec=42.85296375892519, CurrSamplesPerSec=43.52204158465038, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 22:57:27,664 (deepspeed_trainer:226) INFO: 4epoch:train:8901-9000batch: iter_time=1.124e-04, loss_ctc=151.468, loss_att=157.331, acc=0.349, loss=155.676, grad_norm=4.100, loss_scale=1.000, learning_rate=1.698e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-28 22:58:04,922 (deepspeed_trainer:226) INFO: 4epoch:train:9001-9100batch: iter_time=1.070e-04, loss_ctc=132.053, loss_att=130.056, acc=0.381, loss=130.637, grad_norm=3.994, loss_scale=1.000, learning_rate=1.703e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 22:58:43,518 (deepspeed_trainer:226) INFO: 4epoch:train:9101-9200batch: iter_time=1.096e-04, loss_ctc=132.613, loss_att=137.681, acc=0.362, loss=136.127, grad_norm=4.066, loss_scale=1.000, learning_rate=1.708e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-28 22:59:20,880 (deepspeed_trainer:226) INFO: 4epoch:train:9201-9300batch: iter_time=1.118e-04, loss_ctc=145.949, loss_att=143.428, acc=0.366, loss=144.205, grad_norm=4.339, loss_scale=1.000, learning_rate=1.713e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 22:59:53,349 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-28 23:00:20,202 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 23:00:25,828 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 23:00:25,828 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-28 23:00:25,830 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 23:00:53,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:53,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:53,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:53,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:53,811] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:48,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:48,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:54,870] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:55,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,575] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:49,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:00:56,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:41,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:41,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:42,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:42,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:42,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:43,839] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:38,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:38,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:38,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:39,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:39,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:39,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:45,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:39,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:39,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:01:46,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:29,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:29,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:30,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:30,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:31,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:26,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:33,154] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:27,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:27,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:27,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:33,921] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:28,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:29,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:29,197] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:29,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:02:35,852] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:16,485] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:18,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:15,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:21,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:22,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:16,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:16,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:16,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:17,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:24,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:03:20,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 23:04:02,027 (deepspeed_trainer:226) INFO: 4epoch:train:9301-9400batch: iter_time=2.341, loss_ctc=146.766, loss_att=154.512, acc=0.357, loss=152.156, grad_norm=4.287, loss_scale=1.000, learning_rate=1.718e-04, step_time=0.471 [cnode7-008:0/16] 2024-11-28 23:04:39,228 (deepspeed_trainer:226) INFO: 4epoch:train:9401-9500batch: iter_time=1.104e-04, loss_ctc=146.150, loss_att=148.014, acc=0.353, loss=147.455, grad_norm=4.027, loss_scale=1.000, learning_rate=1.723e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 23:05:16,870 (deepspeed_trainer:226) INFO: 4epoch:train:9501-9600batch: iter_time=1.107e-04, loss_ctc=158.952, loss_att=156.387, acc=0.341, loss=157.208, grad_norm=4.323, loss_scale=1.000, learning_rate=1.728e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 23:05:54,719 (deepspeed_trainer:226) INFO: 4epoch:train:9601-9700batch: iter_time=1.110e-04, loss_ctc=143.161, loss_att=139.930, acc=0.375, loss=140.845, grad_norm=3.837, loss_scale=1.000, learning_rate=1.733e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 23:06:32,668 (deepspeed_trainer:226) INFO: 4epoch:train:9701-9800batch: iter_time=1.096e-04, loss_ctc=142.653, loss_att=136.409, acc=0.383, loss=138.260, grad_norm=4.017, loss_scale=1.000, learning_rate=1.738e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-28 23:07:09,811 (deepspeed_trainer:226) INFO: 4epoch:train:9801-9900batch: iter_time=1.103e-04, loss_ctc=134.417, loss_att=127.347, acc=0.382, loss=129.456, grad_norm=3.889, loss_scale=1.000, learning_rate=1.743e-04, step_time=0.371 [2024-11-28 23:07:47,319] [INFO] [logging.py:129:log_dist] [Rank 0] step=55000, skipped=0, lr=[np.float64(0.000175005)], mom=[[0.9, 0.98]] [2024-11-28 23:07:47,320] [INFO] [timer.py:264:stop] epoch=0/micro_step=55000/global_step=55000, RunningAvgSamplesPerSec=42.84261217628952, CurrSamplesPerSec=42.27316183958826, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:07:47,321 (deepspeed_trainer:226) INFO: 4epoch:train:9901-10000batch: iter_time=1.094e-04, loss_ctc=123.127, loss_att=119.449, acc=0.381, loss=120.596, grad_norm=4.370, loss_scale=1.000, learning_rate=1.748e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 23:08:25,194 (deepspeed_trainer:226) INFO: 4epoch:train:10001-10100batch: iter_time=1.076e-04, loss_ctc=142.314, loss_att=143.425, acc=0.377, loss=143.081, grad_norm=4.937, loss_scale=1.000, learning_rate=1.753e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-28 23:09:03,650 (deepspeed_trainer:226) INFO: 4epoch:train:10101-10200batch: iter_time=1.082e-04, loss_ctc=128.940, loss_att=120.531, acc=0.384, loss=123.047, grad_norm=8.202, loss_scale=1.000, learning_rate=1.758e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-28 23:09:41,182 (deepspeed_trainer:226) INFO: 4epoch:train:10201-10300batch: iter_time=1.086e-04, loss_ctc=148.643, loss_att=142.415, acc=0.371, loss=144.273, grad_norm=4.522, loss_scale=1.000, learning_rate=1.763e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 23:10:19,444 (deepspeed_trainer:226) INFO: 4epoch:train:10301-10400batch: iter_time=1.103e-04, loss_ctc=132.912, loss_att=134.689, acc=0.371, loss=134.200, grad_norm=4.104, loss_scale=1.000, learning_rate=1.768e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-28 23:10:56,423 (deepspeed_trainer:226) INFO: 4epoch:train:10401-10500batch: iter_time=1.088e-04, loss_ctc=141.149, loss_att=138.132, acc=0.396, loss=139.037, grad_norm=4.054, loss_scale=1.000, learning_rate=1.773e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:11:33,489 (deepspeed_trainer:226) INFO: 4epoch:train:10501-10600batch: iter_time=1.126e-04, loss_ctc=142.326, loss_att=148.412, acc=0.386, loss=146.548, grad_norm=4.741, loss_scale=1.000, learning_rate=1.778e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 23:12:10,010 (deepspeed_trainer:226) INFO: 4epoch:train:10601-10700batch: iter_time=1.103e-04, loss_ctc=122.556, loss_att=117.659, acc=0.391, loss=119.109, grad_norm=3.684, loss_scale=1.000, learning_rate=1.783e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-28 23:12:46,795 (deepspeed_trainer:226) INFO: 4epoch:train:10701-10800batch: iter_time=1.105e-04, loss_ctc=135.309, loss_att=134.451, acc=0.383, loss=134.694, grad_norm=4.070, loss_scale=1.000, learning_rate=1.788e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 23:13:23,533 (deepspeed_trainer:226) INFO: 4epoch:train:10801-10900batch: iter_time=1.116e-04, loss_ctc=135.639, loss_att=133.849, acc=0.401, loss=134.390, grad_norm=4.254, loss_scale=1.000, learning_rate=1.793e-04, step_time=0.367 [2024-11-28 23:14:00,351] [INFO] [logging.py:129:log_dist] [Rank 0] step=56000, skipped=0, lr=[np.float64(0.000180005)], mom=[[0.9, 0.98]] [2024-11-28 23:14:00,351] [INFO] [timer.py:264:stop] epoch=0/micro_step=56000/global_step=56000, RunningAvgSamplesPerSec=42.85845172143258, CurrSamplesPerSec=42.52620195885506, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:14:00,352 (deepspeed_trainer:226) INFO: 4epoch:train:10901-11000batch: iter_time=1.110e-04, loss_ctc=130.036, loss_att=121.828, acc=0.406, loss=124.265, grad_norm=4.292, loss_scale=1.000, learning_rate=1.798e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-28 23:14:37,290 (deepspeed_trainer:226) INFO: 4epoch:train:11001-11100batch: iter_time=1.099e-04, loss_ctc=121.661, loss_att=121.295, acc=0.400, loss=121.448, grad_norm=6.082, loss_scale=1.000, learning_rate=1.803e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:15:14,992 (deepspeed_trainer:226) INFO: 4epoch:train:11101-11200batch: iter_time=1.103e-04, loss_ctc=147.055, loss_att=141.576, acc=0.389, loss=143.231, grad_norm=4.899, loss_scale=1.000, learning_rate=1.808e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 23:15:36,876 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-28 23:16:03,506 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 23:16:08,902 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 23:16:08,902 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-28 23:16:08,904 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 23:16:36,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:30,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:36,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:30,754] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:37,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:37,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:37,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:31,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:37,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:31,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:31,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:38,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:32,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:32,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:32,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:16:39,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:18,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:18,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:19,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:25,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:25,686] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:19,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:25,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:19,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:26,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:26,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:20,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:20,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:20,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:26,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:27,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:17:33,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:06,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:06,973] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:13,737] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:13,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:07,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:14,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:14,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:08,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:08,224] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:08,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:14,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:08,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:15,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:15,334] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:09,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:25,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:54,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:54,687] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:01,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:02,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:02,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:56,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:56,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:03,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:03,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:56,811] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:56,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:03,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:03,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:18:57,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:00,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:19:13,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 23:19:57,211 (deepspeed_trainer:226) INFO: 4epoch:train:11201-11300batch: iter_time=2.349, loss_ctc=150.512, loss_att=147.719, acc=0.385, loss=148.541, grad_norm=4.776, loss_scale=1.000, learning_rate=1.813e-04, step_time=0.474 [cnode7-008:0/16] 2024-11-28 23:20:35,983 (deepspeed_trainer:226) INFO: 4epoch:train:11301-11400batch: iter_time=1.045e-04, loss_ctc=127.860, loss_att=127.171, acc=0.402, loss=127.349, grad_norm=3.983, loss_scale=1.000, learning_rate=1.818e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-28 23:21:13,274 (deepspeed_trainer:226) INFO: 4epoch:train:11401-11500batch: iter_time=1.078e-04, loss_ctc=150.794, loss_att=148.515, acc=0.390, loss=149.296, grad_norm=4.246, loss_scale=1.000, learning_rate=1.823e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 23:21:50,472 (deepspeed_trainer:226) INFO: 4epoch:train:11501-11600batch: iter_time=1.073e-04, loss_ctc=137.141, loss_att=131.564, acc=0.416, loss=133.251, grad_norm=4.006, loss_scale=1.000, learning_rate=1.828e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 23:22:27,400 (deepspeed_trainer:226) INFO: 4epoch:train:11601-11700batch: iter_time=1.073e-04, loss_ctc=133.851, loss_att=117.534, acc=0.431, loss=122.470, grad_norm=3.768, loss_scale=1.000, learning_rate=1.833e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:23:04,287 (deepspeed_trainer:226) INFO: 4epoch:train:11701-11800batch: iter_time=1.109e-04, loss_ctc=128.651, loss_att=118.972, acc=0.419, loss=121.841, grad_norm=3.849, loss_scale=1.000, learning_rate=1.838e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:23:41,045 (deepspeed_trainer:226) INFO: 4epoch:train:11801-11900batch: iter_time=1.079e-04, loss_ctc=116.203, loss_att=116.259, acc=0.421, loss=116.222, grad_norm=3.633, loss_scale=1.000, learning_rate=1.843e-04, step_time=0.367 [2024-11-28 23:24:18,153] [INFO] [logging.py:129:log_dist] [Rank 0] step=57000, skipped=0, lr=[np.float64(0.000185005)], mom=[[0.9, 0.98]] [2024-11-28 23:24:18,153] [INFO] [timer.py:264:stop] epoch=0/micro_step=57000/global_step=57000, RunningAvgSamplesPerSec=42.85392411714734, CurrSamplesPerSec=43.37024093674384, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:24:18,155 (deepspeed_trainer:226) INFO: 4epoch:train:11901-12000batch: iter_time=1.069e-04, loss_ctc=137.964, loss_att=131.238, acc=0.430, loss=133.240, grad_norm=3.952, loss_scale=1.000, learning_rate=1.848e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 23:24:55,627 (deepspeed_trainer:226) INFO: 4epoch:train:12001-12100batch: iter_time=1.077e-04, loss_ctc=128.653, loss_att=120.325, acc=0.407, loss=122.772, grad_norm=4.130, loss_scale=1.000, learning_rate=1.853e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 23:25:34,070 (deepspeed_trainer:226) INFO: 4epoch:train:12101-12200batch: iter_time=1.065e-04, loss_ctc=133.002, loss_att=126.218, acc=0.422, loss=128.299, grad_norm=4.157, loss_scale=1.000, learning_rate=1.858e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-28 23:26:11,822 (deepspeed_trainer:226) INFO: 4epoch:train:12201-12300batch: iter_time=1.075e-04, loss_ctc=130.734, loss_att=133.768, acc=0.411, loss=132.919, grad_norm=3.771, loss_scale=1.000, learning_rate=1.863e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-28 23:26:50,124 (deepspeed_trainer:226) INFO: 4epoch:train:12301-12400batch: iter_time=1.150e-04, loss_ctc=133.026, loss_att=130.130, acc=0.447, loss=131.059, grad_norm=3.790, loss_scale=1.000, learning_rate=1.868e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-28 23:27:28,375 (deepspeed_trainer:226) INFO: 4epoch:train:12401-12500batch: iter_time=1.109e-04, loss_ctc=131.127, loss_att=133.180, acc=0.419, loss=132.537, grad_norm=4.138, loss_scale=1.000, learning_rate=1.873e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-28 23:28:05,494 (deepspeed_trainer:226) INFO: 4epoch:train:12501-12600batch: iter_time=1.085e-04, loss_ctc=121.121, loss_att=111.654, acc=0.437, loss=114.502, grad_norm=3.991, loss_scale=1.000, learning_rate=1.878e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 23:28:43,375 (deepspeed_trainer:226) INFO: 4epoch:train:12601-12700batch: iter_time=1.102e-04, loss_ctc=132.275, loss_att=125.714, acc=0.430, loss=127.689, grad_norm=3.802, loss_scale=1.000, learning_rate=1.883e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-28 23:29:24,039 (deepspeed_trainer:226) INFO: 4epoch:train:12701-12800batch: iter_time=1.104e-04, loss_ctc=135.118, loss_att=132.945, acc=0.431, loss=133.629, grad_norm=4.078, loss_scale=1.000, learning_rate=1.888e-04, step_time=0.407 [cnode7-008:0/16] 2024-11-28 23:30:00,975 (deepspeed_trainer:226) INFO: 4epoch:train:12801-12900batch: iter_time=1.084e-04, loss_ctc=115.475, loss_att=105.799, acc=0.455, loss=108.696, grad_norm=3.712, loss_scale=1.000, learning_rate=1.893e-04, step_time=0.369 [2024-11-28 23:30:37,866] [INFO] [logging.py:129:log_dist] [Rank 0] step=58000, skipped=0, lr=[np.float64(0.000190005)], mom=[[0.9, 0.98]] [2024-11-28 23:30:37,867] [INFO] [timer.py:264:stop] epoch=0/micro_step=58000/global_step=58000, RunningAvgSamplesPerSec=42.856338684993325, CurrSamplesPerSec=42.408548182233766, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:30:37,868 (deepspeed_trainer:226) INFO: 4epoch:train:12901-13000batch: iter_time=1.118e-04, loss_ctc=114.994, loss_att=105.640, acc=0.457, loss=108.440, grad_norm=3.556, loss_scale=1.000, learning_rate=1.898e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:31:15,131 (deepspeed_trainer:226) INFO: 4epoch:train:13001-13100batch: iter_time=1.077e-04, loss_ctc=149.750, loss_att=141.817, acc=0.430, loss=144.200, grad_norm=4.825, loss_scale=1.000, learning_rate=1.903e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 23:31:27,811 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-28 23:31:54,826 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 23:32:00,415 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 23:32:00,415 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-28 23:32:00,418 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 23:32:27,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:21,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:21,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:21,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:29,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:29,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:23,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:29,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:29,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:23,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:29,824] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:23,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:30,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:23,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:23,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:32:30,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:10,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:16,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:11,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:11,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:11,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:17,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:11,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:18,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:18,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:12,032] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:18,801] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:18,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:19,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:20,154] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:15,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:15,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:57,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:05,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:58,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:59,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:06,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:06,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:33:59,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:06,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:00,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:07,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:07,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:00,710] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:07,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:08,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:06,270] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:06,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:44,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:52,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:46,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:53,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:53,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:53,754] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:47,236] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:47,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:54,816] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:55,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:55,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:48,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:56,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:52,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:56,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:34:56,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 23:35:56,623 (deepspeed_trainer:226) INFO: 4epoch:train:13101-13200batch: iter_time=2.374, loss_ctc=133.129, loss_att=130.458, acc=0.421, loss=131.299, grad_norm=4.533, loss_scale=1.000, learning_rate=1.908e-04, step_time=0.441 [cnode7-008:0/16] 2024-11-28 23:36:34,344 (deepspeed_trainer:226) INFO: 4epoch:train:13201-13300batch: iter_time=1.048e-04, loss_ctc=126.647, loss_att=116.250, acc=0.448, loss=119.347, grad_norm=3.820, loss_scale=1.000, learning_rate=1.913e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-28 23:37:11,476 (deepspeed_trainer:226) INFO: 4epoch:train:13301-13400batch: iter_time=1.081e-04, loss_ctc=147.877, loss_att=141.532, acc=0.416, loss=143.481, grad_norm=4.493, loss_scale=1.000, learning_rate=1.918e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-28 23:37:49,576 (deepspeed_trainer:226) INFO: 4epoch:train:13401-13500batch: iter_time=1.062e-04, loss_ctc=130.162, loss_att=112.309, acc=0.464, loss=117.695, grad_norm=3.947, loss_scale=1.000, learning_rate=1.923e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-28 23:38:26,277 (deepspeed_trainer:226) INFO: 4epoch:train:13501-13600batch: iter_time=1.070e-04, loss_ctc=128.398, loss_att=107.461, acc=0.466, loss=113.700, grad_norm=3.856, loss_scale=1.000, learning_rate=1.928e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 23:39:03,319 (deepspeed_trainer:226) INFO: 4epoch:train:13601-13700batch: iter_time=1.122e-04, loss_ctc=116.686, loss_att=101.064, acc=0.467, loss=105.772, grad_norm=3.441, loss_scale=1.000, learning_rate=1.933e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 23:39:42,508 (deepspeed_trainer:226) INFO: 4epoch:train:13701-13800batch: iter_time=1.135e-04, loss_ctc=116.671, loss_att=109.049, acc=0.447, loss=111.320, grad_norm=3.808, loss_scale=1.000, learning_rate=1.938e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-28 23:40:21,207 (deepspeed_trainer:226) INFO: 4epoch:train:13801-13900batch: iter_time=1.112e-04, loss_ctc=133.801, loss_att=120.114, acc=0.464, loss=124.213, grad_norm=3.753, loss_scale=1.000, learning_rate=1.943e-04, step_time=0.386 [2024-11-28 23:40:57,835] [INFO] [logging.py:129:log_dist] [Rank 0] step=59000, skipped=0, lr=[np.float64(0.000195005)], mom=[[0.9, 0.98]] [2024-11-28 23:40:57,836] [INFO] [timer.py:264:stop] epoch=0/micro_step=59000/global_step=59000, RunningAvgSamplesPerSec=42.85236227510021, CurrSamplesPerSec=44.64946169863195, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:40:57,837 (deepspeed_trainer:226) INFO: 4epoch:train:13901-14000batch: iter_time=1.115e-04, loss_ctc=134.377, loss_att=121.668, acc=0.441, loss=125.546, grad_norm=4.455, loss_scale=1.000, learning_rate=1.948e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 23:41:34,440 (deepspeed_trainer:226) INFO: 4epoch:train:14001-14100batch: iter_time=1.137e-04, loss_ctc=124.326, loss_att=116.657, acc=0.437, loss=118.978, grad_norm=3.732, loss_scale=1.000, learning_rate=1.953e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 23:42:11,133 (deepspeed_trainer:226) INFO: 4epoch:train:14101-14200batch: iter_time=1.131e-04, loss_ctc=124.345, loss_att=115.473, acc=0.469, loss=118.179, grad_norm=3.444, loss_scale=1.000, learning_rate=1.958e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-28 23:42:48,058 (deepspeed_trainer:226) INFO: 4epoch:train:14201-14300batch: iter_time=1.129e-04, loss_ctc=129.790, loss_att=122.771, acc=0.455, loss=124.851, grad_norm=4.274, loss_scale=1.000, learning_rate=1.963e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-28 23:43:24,633 (deepspeed_trainer:226) INFO: 4epoch:train:14301-14400batch: iter_time=1.076e-04, loss_ctc=123.671, loss_att=119.156, acc=0.457, loss=120.502, grad_norm=3.672, loss_scale=1.000, learning_rate=1.968e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-28 23:44:01,353 (deepspeed_trainer:226) INFO: 4epoch:train:14401-14500batch: iter_time=1.074e-04, loss_ctc=118.411, loss_att=105.954, acc=0.470, loss=109.683, grad_norm=4.115, loss_scale=1.000, learning_rate=1.973e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-28 23:44:38,805 (deepspeed_trainer:226) INFO: 4epoch:train:14501-14600batch: iter_time=1.081e-04, loss_ctc=129.599, loss_att=118.929, acc=0.455, loss=122.111, grad_norm=4.040, loss_scale=1.000, learning_rate=1.978e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 23:45:15,852 (deepspeed_trainer:226) INFO: 4epoch:train:14601-14700batch: iter_time=1.083e-04, loss_ctc=122.570, loss_att=111.762, acc=0.474, loss=114.952, grad_norm=4.119, loss_scale=1.000, learning_rate=1.983e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 23:45:52,923 (deepspeed_trainer:226) INFO: 4epoch:train:14701-14800batch: iter_time=1.120e-04, loss_ctc=114.772, loss_att=105.480, acc=0.466, loss=108.270, grad_norm=3.816, loss_scale=1.000, learning_rate=1.988e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-28 23:46:29,578 (deepspeed_trainer:226) INFO: 4epoch:train:14801-14900batch: iter_time=1.089e-04, loss_ctc=116.149, loss_att=101.994, acc=0.484, loss=106.291, grad_norm=3.840, loss_scale=1.000, learning_rate=1.993e-04, step_time=0.366 [2024-11-28 23:47:07,780] [INFO] [logging.py:129:log_dist] [Rank 0] step=60000, skipped=0, lr=[np.float64(0.0001999983333541664)], mom=[[0.9, 0.98]] [2024-11-28 23:47:07,781] [INFO] [timer.py:264:stop] epoch=0/micro_step=60000/global_step=60000, RunningAvgSamplesPerSec=42.8725926647815, CurrSamplesPerSec=47.275404404415426, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:47:07,782 (deepspeed_trainer:226) INFO: 4epoch:train:14901-15000batch: iter_time=1.060e-04, loss_ctc=143.273, loss_att=136.585, acc=0.442, loss=138.505, grad_norm=4.309, loss_scale=1.000, learning_rate=1.998e-04, step_time=0.382 [2024-11-28 23:47:22,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:23,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:23,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:23,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:16,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:17,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:17,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:23,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:23,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:24,037] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:24,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:38,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:38,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:38,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:32,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:39,322] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:39,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:32,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:33,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:33,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:40,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:40,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:40,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:34,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:34,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:34,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:34,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:53,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:54,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:47,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:54,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:55,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:48,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:55,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:48,784] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:48,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:55,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:49,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:56,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:56,143] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:49,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:49,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:47:50,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:09,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:09,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:02,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:09,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:10,488] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:11,078] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:11,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:04,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:04,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:11,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:04,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:11,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:05,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:05,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:05,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:05,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:48:25,060] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 4 is about to be saved! [2024-11-28 23:48:25,091] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/mp_rank_00_model_states.pt [2024-11-28 23:48:25,091] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/mp_rank_00_model_states.pt... [2024-11-28 23:48:26,997] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/mp_rank_00_model_states.pt. [2024-11-28 23:48:27,156] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,422] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,159] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,424] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,425] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,162] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,162] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,426] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,427] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,429] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,429] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-28 23:48:20,429] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-28 23:48:27,755] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,755] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,755] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,763] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,763] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,763] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,809] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,809] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,809] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,086] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,087] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,087] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,089] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,090] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,090] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,149] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,150] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,150] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,164] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,164] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,164] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,169] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,170] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,170] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,912] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,912] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,912] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,915] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,915] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,915] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,930] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,930] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,930] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,963] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,968] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,968] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:27,970] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-28 23:48:27,970] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-28 23:48:27,971] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,258] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,259] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,259] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,267] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,267] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,267] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [2024-11-28 23:48:21,289] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-28 23:48:21,289] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_4/4/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-28 23:48:21,289] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 4 is ready now! [cnode7-008:0/16] 2024-11-28 23:48:28,030 (deepspeed_trainer:158) INFO: 4epoch results: [train] iter_time=0.125, loss_ctc=153.799, loss_att=148.722, acc=0.348, loss=150.245, grad_norm=4.058, loss_scale=1.000, learning_rate=1.625e-04, step_time=0.378, time=2 hours, 6 minutes and 10.39 seconds, total_count=60004, gpu_max_cached_mem_GB=27.396, [valid] loss_ctc=25.625, cer_ctc=0.617, loss_att=34.000, acc=0.466, cer=0.604, wer=1.000, loss=31.500, time=1 minute and 8.56 seconds, total_count=4, gpu_max_cached_mem_GB=27.396 [cnode7-008:0/16] 2024-11-28 23:48:30,373 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-28 23:48:56,243 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-28 23:49:01,625 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-28 23:49:01,625 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-28 23:49:01,627 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-28 23:49:22,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:22,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:15,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:16,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:16,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:16,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:17,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:17,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:17,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:24,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:24,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:25,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:25,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:19,532] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:26,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:49:26,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:10,037] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:03,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:03,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:10,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:04,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:04,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:05,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:12,971] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:06,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:13,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:13,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:14,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:14,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:08,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:08,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:15,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:57,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:58,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:53,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:53,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:54,971] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:00,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:01,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:55,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:56,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:02,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:02,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:02,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:57,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:50:57,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:04,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:01,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:45,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:46,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:41,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:41,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:47,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:43,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:49,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:49,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:50,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:50,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:45,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:45,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:52,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:47,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:48,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-28 23:51:51,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-28 23:52:59,487 (deepspeed_trainer:226) INFO: 5epoch:train:1-100batch: iter_time=2.253, loss_ctc=125.096, loss_att=112.534, acc=0.459, loss=116.301, grad_norm=4.158, loss_scale=1.000, learning_rate=1.999e-04, step_time=0.439 [cnode7-008:0/16] 2024-11-28 23:53:36,913 (deepspeed_trainer:226) INFO: 5epoch:train:101-200batch: iter_time=1.035e-04, loss_ctc=119.689, loss_att=111.147, acc=0.467, loss=113.733, grad_norm=3.746, loss_scale=1.000, learning_rate=1.997e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 23:54:14,476 (deepspeed_trainer:226) INFO: 5epoch:train:201-300batch: iter_time=1.084e-04, loss_ctc=135.604, loss_att=131.499, acc=0.465, loss=132.771, grad_norm=4.170, loss_scale=1.000, learning_rate=1.996e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-28 23:54:51,714 (deepspeed_trainer:226) INFO: 5epoch:train:301-400batch: iter_time=1.096e-04, loss_ctc=139.471, loss_att=135.861, acc=0.453, loss=136.949, grad_norm=4.693, loss_scale=1.000, learning_rate=1.994e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 23:55:29,143 (deepspeed_trainer:226) INFO: 5epoch:train:401-500batch: iter_time=1.286e-04, loss_ctc=124.619, loss_att=111.954, acc=0.453, loss=115.741, grad_norm=3.836, loss_scale=1.000, learning_rate=1.993e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-28 23:56:06,399 (deepspeed_trainer:226) INFO: 5epoch:train:501-600batch: iter_time=1.113e-04, loss_ctc=133.698, loss_att=117.034, acc=0.488, loss=122.096, grad_norm=4.003, loss_scale=1.000, learning_rate=1.991e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-28 23:56:44,140 (deepspeed_trainer:226) INFO: 5epoch:train:601-700batch: iter_time=1.068e-04, loss_ctc=135.936, loss_att=128.094, acc=0.473, loss=130.406, grad_norm=4.064, loss_scale=1.000, learning_rate=1.989e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 23:57:21,864 (deepspeed_trainer:226) INFO: 5epoch:train:701-800batch: iter_time=1.114e-04, loss_ctc=119.870, loss_att=108.941, acc=0.466, loss=112.284, grad_norm=3.998, loss_scale=1.000, learning_rate=1.988e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-28 23:57:58,791 (deepspeed_trainer:226) INFO: 5epoch:train:801-900batch: iter_time=1.095e-04, loss_ctc=123.721, loss_att=116.200, acc=0.472, loss=118.471, grad_norm=4.294, loss_scale=1.000, learning_rate=1.986e-04, step_time=0.369 [2024-11-28 23:58:36,163] [INFO] [logging.py:129:log_dist] [Rank 0] step=61000, skipped=0, lr=[np.float64(0.00019835225564068812)], mom=[[0.9, 0.98]] [2024-11-28 23:58:36,164] [INFO] [timer.py:264:stop] epoch=0/micro_step=61000/global_step=61000, RunningAvgSamplesPerSec=42.87240681261014, CurrSamplesPerSec=44.74445705730537, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-28 23:58:36,165 (deepspeed_trainer:226) INFO: 5epoch:train:901-1000batch: iter_time=1.130e-04, loss_ctc=132.757, loss_att=117.683, acc=0.480, loss=122.224, grad_norm=3.913, loss_scale=1.000, learning_rate=1.984e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-28 23:59:15,302 (deepspeed_trainer:226) INFO: 5epoch:train:1001-1100batch: iter_time=1.205e-04, loss_ctc=123.049, loss_att=119.831, acc=0.468, loss=120.824, grad_norm=4.134, loss_scale=1.000, learning_rate=1.983e-04, step_time=0.391 [cnode7-008:0/16] 2024-11-28 23:59:54,559 (deepspeed_trainer:226) INFO: 5epoch:train:1101-1200batch: iter_time=1.082e-04, loss_ctc=119.360, loss_att=108.193, acc=0.460, loss=111.558, grad_norm=4.045, loss_scale=1.000, learning_rate=1.981e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 00:00:31,887 (deepspeed_trainer:226) INFO: 5epoch:train:1201-1300batch: iter_time=1.091e-04, loss_ctc=125.537, loss_att=120.473, acc=0.474, loss=121.966, grad_norm=3.602, loss_scale=1.000, learning_rate=1.979e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:01:09,028 (deepspeed_trainer:226) INFO: 5epoch:train:1301-1400batch: iter_time=1.074e-04, loss_ctc=135.773, loss_att=111.310, acc=0.502, loss=118.678, grad_norm=3.488, loss_scale=1.000, learning_rate=1.978e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:01:46,234 (deepspeed_trainer:226) INFO: 5epoch:train:1401-1500batch: iter_time=1.071e-04, loss_ctc=132.490, loss_att=123.589, acc=0.468, loss=126.275, grad_norm=3.673, loss_scale=1.000, learning_rate=1.976e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 00:02:23,412 (deepspeed_trainer:226) INFO: 5epoch:train:1501-1600batch: iter_time=1.069e-04, loss_ctc=129.151, loss_att=115.101, acc=0.482, loss=119.323, grad_norm=3.935, loss_scale=1.000, learning_rate=1.975e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:03:00,362 (deepspeed_trainer:226) INFO: 5epoch:train:1601-1700batch: iter_time=1.085e-04, loss_ctc=127.278, loss_att=119.062, acc=0.465, loss=121.528, grad_norm=3.979, loss_scale=1.000, learning_rate=1.973e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 00:03:37,464 (deepspeed_trainer:226) INFO: 5epoch:train:1701-1800batch: iter_time=1.069e-04, loss_ctc=126.647, loss_att=111.448, acc=0.474, loss=115.979, grad_norm=3.756, loss_scale=1.000, learning_rate=1.971e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:04:09,874 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-29 00:04:37,065 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 00:04:42,991 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 00:04:42,991 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-29 00:04:42,994 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 00:05:13,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:08,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,529] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:14,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:57,908] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:58,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:58,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:04,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:04,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:05,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:05,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:05:59,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:05,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:05,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:00,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:00,078] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:00,156] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:00,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:06,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:07,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:52,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:52,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:46,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:52,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:47,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:47,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:47,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:48,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:53,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:54,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:48,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:48,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:49,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:55,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:56,157] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:06:57,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:40,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:34,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:34,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:41,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:41,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:35,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:36,218] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:36,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:36,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:42,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:42,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:37,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:38,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:44,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:45,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:07:46,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 00:08:23,604 (deepspeed_trainer:226) INFO: 5epoch:train:1801-1900batch: iter_time=2.453, loss_ctc=122.354, loss_att=108.334, acc=0.490, loss=112.556, grad_norm=3.890, loss_scale=1.000, learning_rate=1.970e-04, step_time=0.408 [2024-11-29 00:09:01,970] [INFO] [logging.py:129:log_dist] [Rank 0] step=62000, skipped=0, lr=[np.float64(0.0001967461640789701)], mom=[[0.9, 0.98]] [2024-11-29 00:09:01,971] [INFO] [timer.py:264:stop] epoch=0/micro_step=62000/global_step=62000, RunningAvgSamplesPerSec=42.87245557369638, CurrSamplesPerSec=46.42046199926025, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:09:01,972 (deepspeed_trainer:226) INFO: 5epoch:train:1901-2000batch: iter_time=1.094e-04, loss_ctc=113.407, loss_att=100.079, acc=0.492, loss=104.089, grad_norm=3.380, loss_scale=1.000, learning_rate=1.968e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 00:09:41,488 (deepspeed_trainer:226) INFO: 5epoch:train:2001-2100batch: iter_time=1.048e-04, loss_ctc=107.804, loss_att=101.231, acc=0.487, loss=103.199, grad_norm=3.786, loss_scale=1.000, learning_rate=1.967e-04, step_time=0.395 [cnode7-008:0/16] 2024-11-29 00:10:19,105 (deepspeed_trainer:226) INFO: 5epoch:train:2101-2200batch: iter_time=1.055e-04, loss_ctc=146.759, loss_att=131.407, acc=0.494, loss=136.047, grad_norm=4.075, loss_scale=1.000, learning_rate=1.965e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 00:10:56,106 (deepspeed_trainer:226) INFO: 5epoch:train:2201-2300batch: iter_time=1.128e-04, loss_ctc=117.972, loss_att=109.785, acc=0.470, loss=112.188, grad_norm=4.058, loss_scale=1.000, learning_rate=1.963e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:11:32,929 (deepspeed_trainer:226) INFO: 5epoch:train:2301-2400batch: iter_time=1.150e-04, loss_ctc=118.776, loss_att=106.563, acc=0.476, loss=110.241, grad_norm=3.724, loss_scale=1.000, learning_rate=1.962e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 00:12:10,263 (deepspeed_trainer:226) INFO: 5epoch:train:2401-2500batch: iter_time=1.238e-04, loss_ctc=131.257, loss_att=106.817, acc=0.515, loss=114.158, grad_norm=3.912, loss_scale=1.000, learning_rate=1.960e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:12:47,651 (deepspeed_trainer:226) INFO: 5epoch:train:2501-2600batch: iter_time=1.156e-04, loss_ctc=146.331, loss_att=136.521, acc=0.476, loss=139.468, grad_norm=4.022, loss_scale=1.000, learning_rate=1.959e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 00:13:24,641 (deepspeed_trainer:226) INFO: 5epoch:train:2601-2700batch: iter_time=1.168e-04, loss_ctc=101.373, loss_att=92.061, acc=0.491, loss=94.819, grad_norm=3.720, loss_scale=1.000, learning_rate=1.957e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 00:14:01,857 (deepspeed_trainer:226) INFO: 5epoch:train:2701-2800batch: iter_time=1.229e-04, loss_ctc=116.304, loss_att=99.338, acc=0.497, loss=104.388, grad_norm=3.642, loss_scale=1.000, learning_rate=1.956e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 00:14:40,069 (deepspeed_trainer:226) INFO: 5epoch:train:2801-2900batch: iter_time=1.217e-04, loss_ctc=132.394, loss_att=118.390, acc=0.493, loss=122.611, grad_norm=3.948, loss_scale=1.000, learning_rate=1.954e-04, step_time=0.382 [2024-11-29 00:15:17,430] [INFO] [logging.py:129:log_dist] [Rank 0] step=63000, skipped=0, lr=[np.float64(0.0001951784655604126)], mom=[[0.9, 0.98]] [2024-11-29 00:15:17,431] [INFO] [timer.py:264:stop] epoch=0/micro_step=63000/global_step=63000, RunningAvgSamplesPerSec=42.88165417060372, CurrSamplesPerSec=46.179387049292835, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:15:17,432 (deepspeed_trainer:226) INFO: 5epoch:train:2901-3000batch: iter_time=1.163e-04, loss_ctc=114.870, loss_att=105.108, acc=0.485, loss=108.097, grad_norm=3.739, loss_scale=1.000, learning_rate=1.953e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:15:54,246 (deepspeed_trainer:226) INFO: 5epoch:train:3001-3100batch: iter_time=1.177e-04, loss_ctc=116.700, loss_att=106.688, acc=0.475, loss=109.684, grad_norm=4.050, loss_scale=1.000, learning_rate=1.951e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 00:16:33,534 (deepspeed_trainer:226) INFO: 5epoch:train:3101-3200batch: iter_time=1.164e-04, loss_ctc=129.540, loss_att=116.801, acc=0.497, loss=120.659, grad_norm=3.754, loss_scale=1.000, learning_rate=1.949e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 00:17:12,340 (deepspeed_trainer:226) INFO: 5epoch:train:3201-3300batch: iter_time=1.214e-04, loss_ctc=117.304, loss_att=92.353, acc=0.529, loss=99.853, grad_norm=3.263, loss_scale=1.000, learning_rate=1.948e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 00:17:50,759 (deepspeed_trainer:226) INFO: 5epoch:train:3301-3400batch: iter_time=1.303e-04, loss_ctc=137.035, loss_att=124.980, acc=0.479, loss=128.577, grad_norm=3.968, loss_scale=1.000, learning_rate=1.946e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 00:18:27,827 (deepspeed_trainer:226) INFO: 5epoch:train:3401-3500batch: iter_time=1.220e-04, loss_ctc=122.315, loss_att=109.338, acc=0.507, loss=113.192, grad_norm=3.532, loss_scale=1.000, learning_rate=1.945e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:19:04,213 (deepspeed_trainer:226) INFO: 5epoch:train:3501-3600batch: iter_time=1.240e-04, loss_ctc=121.037, loss_att=103.160, acc=0.480, loss=108.494, grad_norm=4.406, loss_scale=1.000, learning_rate=1.943e-04, step_time=0.363 [cnode7-008:0/16] 2024-11-29 00:19:40,703 (deepspeed_trainer:226) INFO: 5epoch:train:3601-3700batch: iter_time=1.278e-04, loss_ctc=113.654, loss_att=96.719, acc=0.491, loss=101.766, grad_norm=3.845, loss_scale=1.000, learning_rate=1.942e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 00:20:02,701 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-29 00:20:30,303 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 00:20:36,155 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 00:20:36,156 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-29 00:20:36,158 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 00:20:54,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:55,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:56,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:56,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:56,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:57,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:57,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:20:57,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:04,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:04,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:04,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:05,224] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:05,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:05,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:05,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:05,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:43,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:43,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:43,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:44,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:44,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:44,540] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:44,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:53,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:54,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:54,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:54,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:55,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:55,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:55,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:49,839] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:21:56,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:30,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:31,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:31,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:33,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:33,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:33,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:41,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:42,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:36,213] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:42,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:43,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:43,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:43,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:43,815] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:39,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:22:48,360] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:17,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:18,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:19,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:19,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:20,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:20,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:28,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:29,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:29,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:30,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:30,629] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:31,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:26,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:34,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:28,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:23:37,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 00:24:22,163 (deepspeed_trainer:226) INFO: 5epoch:train:3701-3800batch: iter_time=2.382, loss_ctc=124.172, loss_att=111.565, acc=0.502, loss=115.337, grad_norm=3.842, loss_scale=1.000, learning_rate=1.940e-04, step_time=0.433 [cnode7-008:0/16] 2024-11-29 00:24:59,862 (deepspeed_trainer:226) INFO: 5epoch:train:3801-3900batch: iter_time=1.047e-04, loss_ctc=112.192, loss_att=94.150, acc=0.515, loss=99.578, grad_norm=3.670, loss_scale=1.000, learning_rate=1.939e-04, step_time=0.377 [2024-11-29 00:25:37,684] [INFO] [logging.py:129:log_dist] [Rank 0] step=64000, skipped=0, lr=[np.float64(0.0001936476544439801)], mom=[[0.9, 0.98]] [2024-11-29 00:25:37,685] [INFO] [timer.py:264:stop] epoch=0/micro_step=64000/global_step=64000, RunningAvgSamplesPerSec=42.87858411162093, CurrSamplesPerSec=40.05037936493082, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:25:37,686 (deepspeed_trainer:226) INFO: 5epoch:train:3901-4000batch: iter_time=1.058e-04, loss_ctc=108.826, loss_att=99.391, acc=0.512, loss=102.222, grad_norm=3.532, loss_scale=1.000, learning_rate=1.937e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 00:26:17,345 (deepspeed_trainer:226) INFO: 5epoch:train:4001-4100batch: iter_time=1.053e-04, loss_ctc=140.809, loss_att=124.236, acc=0.507, loss=129.182, grad_norm=3.920, loss_scale=1.000, learning_rate=1.936e-04, step_time=0.396 [cnode7-008:0/16] 2024-11-29 00:26:56,101 (deepspeed_trainer:226) INFO: 5epoch:train:4101-4200batch: iter_time=1.065e-04, loss_ctc=104.959, loss_att=96.581, acc=0.490, loss=99.093, grad_norm=3.839, loss_scale=1.000, learning_rate=1.934e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 00:27:34,533 (deepspeed_trainer:226) INFO: 5epoch:train:4201-4300batch: iter_time=1.049e-04, loss_ctc=131.681, loss_att=112.899, acc=0.503, loss=118.517, grad_norm=4.570, loss_scale=1.000, learning_rate=1.933e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 00:28:11,874 (deepspeed_trainer:226) INFO: 5epoch:train:4301-4400batch: iter_time=1.046e-04, loss_ctc=123.236, loss_att=103.480, acc=0.515, loss=109.410, grad_norm=4.131, loss_scale=1.000, learning_rate=1.931e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:28:48,660 (deepspeed_trainer:226) INFO: 5epoch:train:4401-4500batch: iter_time=1.055e-04, loss_ctc=123.162, loss_att=111.472, acc=0.507, loss=114.976, grad_norm=3.816, loss_scale=1.000, learning_rate=1.930e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 00:29:25,244 (deepspeed_trainer:226) INFO: 5epoch:train:4501-4600batch: iter_time=1.054e-04, loss_ctc=103.879, loss_att=95.937, acc=0.504, loss=98.311, grad_norm=3.404, loss_scale=1.000, learning_rate=1.928e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 00:30:01,837 (deepspeed_trainer:226) INFO: 5epoch:train:4601-4700batch: iter_time=1.064e-04, loss_ctc=123.475, loss_att=100.571, acc=0.511, loss=107.492, grad_norm=3.837, loss_scale=1.000, learning_rate=1.927e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 00:30:38,361 (deepspeed_trainer:226) INFO: 5epoch:train:4701-4800batch: iter_time=1.061e-04, loss_ctc=110.380, loss_att=96.219, acc=0.513, loss=100.478, grad_norm=3.762, loss_scale=1.000, learning_rate=1.925e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 00:31:14,851 (deepspeed_trainer:226) INFO: 5epoch:train:4801-4900batch: iter_time=1.059e-04, loss_ctc=113.698, loss_att=100.946, acc=0.503, loss=104.753, grad_norm=4.257, loss_scale=1.000, learning_rate=1.924e-04, step_time=0.365 [2024-11-29 00:31:51,602] [INFO] [logging.py:129:log_dist] [Rank 0] step=65000, skipped=0, lr=[np.float64(0.0001921523064771243)], mom=[[0.9, 0.98]] [2024-11-29 00:31:51,603] [INFO] [timer.py:264:stop] epoch=0/micro_step=65000/global_step=65000, RunningAvgSamplesPerSec=42.8899132323546, CurrSamplesPerSec=42.201041492530486, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:31:51,604 (deepspeed_trainer:226) INFO: 5epoch:train:4901-5000batch: iter_time=1.056e-04, loss_ctc=117.192, loss_att=108.959, acc=0.495, loss=111.398, grad_norm=4.030, loss_scale=1.000, learning_rate=1.922e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 00:32:29,198 (deepspeed_trainer:226) INFO: 5epoch:train:5001-5100batch: iter_time=1.059e-04, loss_ctc=129.699, loss_att=112.824, acc=0.519, loss=117.907, grad_norm=3.603, loss_scale=1.000, learning_rate=1.921e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 00:33:07,104 (deepspeed_trainer:226) INFO: 5epoch:train:5101-5200batch: iter_time=1.059e-04, loss_ctc=113.803, loss_att=88.955, acc=0.530, loss=96.403, grad_norm=3.499, loss_scale=1.000, learning_rate=1.919e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 00:33:43,695 (deepspeed_trainer:226) INFO: 5epoch:train:5201-5300batch: iter_time=1.055e-04, loss_ctc=131.577, loss_att=117.691, acc=0.497, loss=121.832, grad_norm=4.124, loss_scale=1.000, learning_rate=1.918e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 00:34:20,983 (deepspeed_trainer:226) INFO: 5epoch:train:5301-5400batch: iter_time=1.060e-04, loss_ctc=120.919, loss_att=109.149, acc=0.509, loss=112.656, grad_norm=4.150, loss_scale=1.000, learning_rate=1.916e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:35:00,205 (deepspeed_trainer:226) INFO: 5epoch:train:5401-5500batch: iter_time=1.069e-04, loss_ctc=117.104, loss_att=97.739, acc=0.510, loss=103.549, grad_norm=3.515, loss_scale=1.000, learning_rate=1.915e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 00:35:38,187 (deepspeed_trainer:226) INFO: 5epoch:train:5501-5600batch: iter_time=1.057e-04, loss_ctc=114.670, loss_att=98.048, acc=0.509, loss=103.046, grad_norm=3.937, loss_scale=1.000, learning_rate=1.913e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 00:35:50,476 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-29 00:36:17,845 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 00:36:24,595 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 00:36:24,595 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-29 00:36:24,598 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 00:36:43,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:50,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:44,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:51,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:44,852] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:44,920] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:51,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:44,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:51,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:52,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:45,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:45,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:45,873] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:52,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:52,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:36:53,296] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:34,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:34,237] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:34,448] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:34,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:34,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:41,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:41,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:35,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:35,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:35,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:42,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:42,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:42,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:42,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:43,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:37:43,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:22,648] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:22,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:23,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:23,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:31,032] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:24,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:24,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:24,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:24,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:31,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:32,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:32,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:32,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:33,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:33,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:38:33,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:17,754] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:11,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:11,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:11,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:12,270] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:12,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:19,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:13,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:13,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:19,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:20,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:20,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:21,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:14,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:21,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:39:22,280] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 00:40:16,182 (deepspeed_trainer:226) INFO: 5epoch:train:5601-5700batch: iter_time=2.371, loss_ctc=111.667, loss_att=97.847, acc=0.516, loss=101.951, grad_norm=3.767, loss_scale=1.000, learning_rate=1.912e-04, step_time=0.408 [cnode7-008:0/16] 2024-11-29 00:40:52,878 (deepspeed_trainer:226) INFO: 5epoch:train:5701-5800batch: iter_time=1.058e-04, loss_ctc=107.794, loss_att=96.874, acc=0.520, loss=100.138, grad_norm=3.512, loss_scale=1.000, learning_rate=1.911e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 00:41:29,958 (deepspeed_trainer:226) INFO: 5epoch:train:5801-5900batch: iter_time=1.067e-04, loss_ctc=117.852, loss_att=111.531, acc=0.522, loss=113.457, grad_norm=3.321, loss_scale=1.000, learning_rate=1.909e-04, step_time=0.370 [2024-11-29 00:42:07,024] [INFO] [logging.py:129:log_dist] [Rank 0] step=66000, skipped=0, lr=[np.float64(0.00019069107322524788)], mom=[[0.9, 0.98]] [2024-11-29 00:42:07,025] [INFO] [timer.py:264:stop] epoch=0/micro_step=66000/global_step=66000, RunningAvgSamplesPerSec=42.89339901129843, CurrSamplesPerSec=44.916567456673135, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:42:07,026 (deepspeed_trainer:226) INFO: 5epoch:train:5901-6000batch: iter_time=1.069e-04, loss_ctc=129.422, loss_att=119.303, acc=0.518, loss=122.353, grad_norm=4.287, loss_scale=1.000, learning_rate=1.908e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:42:44,964 (deepspeed_trainer:226) INFO: 5epoch:train:6001-6100batch: iter_time=1.083e-04, loss_ctc=107.808, loss_att=94.819, acc=0.508, loss=98.689, grad_norm=3.892, loss_scale=1.000, learning_rate=1.906e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 00:43:22,044 (deepspeed_trainer:226) INFO: 5epoch:train:6101-6200batch: iter_time=1.084e-04, loss_ctc=123.440, loss_att=112.082, acc=0.515, loss=115.475, grad_norm=3.996, loss_scale=1.000, learning_rate=1.905e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:43:59,117 (deepspeed_trainer:226) INFO: 5epoch:train:6201-6300batch: iter_time=1.096e-04, loss_ctc=124.236, loss_att=103.926, acc=0.540, loss=110.011, grad_norm=3.816, loss_scale=1.000, learning_rate=1.903e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:44:39,153 (deepspeed_trainer:226) INFO: 5epoch:train:6301-6400batch: iter_time=1.093e-04, loss_ctc=113.871, loss_att=107.916, acc=0.524, loss=109.713, grad_norm=4.237, loss_scale=1.000, learning_rate=1.902e-04, step_time=0.400 [cnode7-008:0/16] 2024-11-29 00:45:17,214 (deepspeed_trainer:226) INFO: 5epoch:train:6401-6500batch: iter_time=1.073e-04, loss_ctc=103.514, loss_att=94.638, acc=0.521, loss=97.297, grad_norm=3.743, loss_scale=1.000, learning_rate=1.900e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 00:45:55,036 (deepspeed_trainer:226) INFO: 5epoch:train:6501-6600batch: iter_time=1.075e-04, loss_ctc=118.906, loss_att=104.422, acc=0.526, loss=108.812, grad_norm=3.900, loss_scale=1.000, learning_rate=1.899e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 00:46:32,046 (deepspeed_trainer:226) INFO: 5epoch:train:6601-6700batch: iter_time=1.100e-04, loss_ctc=111.388, loss_att=100.483, acc=0.527, loss=103.769, grad_norm=4.449, loss_scale=1.000, learning_rate=1.898e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:47:08,857 (deepspeed_trainer:226) INFO: 5epoch:train:6701-6800batch: iter_time=1.083e-04, loss_ctc=107.873, loss_att=91.107, acc=0.515, loss=96.158, grad_norm=4.511, loss_scale=1.000, learning_rate=1.896e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 00:47:46,663 (deepspeed_trainer:226) INFO: 5epoch:train:6801-6900batch: iter_time=1.060e-04, loss_ctc=116.750, loss_att=112.327, acc=0.515, loss=113.659, grad_norm=3.904, loss_scale=1.000, learning_rate=1.895e-04, step_time=0.378 [2024-11-29 00:48:24,206] [INFO] [logging.py:129:log_dist] [Rank 0] step=67000, skipped=0, lr=[np.float64(0.00018926267696018666)], mom=[[0.9, 0.98]] [2024-11-29 00:48:24,207] [INFO] [timer.py:264:stop] epoch=0/micro_step=67000/global_step=67000, RunningAvgSamplesPerSec=42.898800836391565, CurrSamplesPerSec=43.823586963910564, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:48:24,208 (deepspeed_trainer:226) INFO: 5epoch:train:6901-7000batch: iter_time=1.090e-04, loss_ctc=127.168, loss_att=105.350, acc=0.536, loss=111.883, grad_norm=3.690, loss_scale=1.000, learning_rate=1.893e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 00:49:01,575 (deepspeed_trainer:226) INFO: 5epoch:train:7001-7100batch: iter_time=1.079e-04, loss_ctc=113.229, loss_att=94.080, acc=0.539, loss=99.853, grad_norm=3.168, loss_scale=1.000, learning_rate=1.892e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:49:39,013 (deepspeed_trainer:226) INFO: 5epoch:train:7101-7200batch: iter_time=1.110e-04, loss_ctc=117.923, loss_att=107.739, acc=0.523, loss=110.789, grad_norm=3.818, loss_scale=1.000, learning_rate=1.891e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 00:50:16,168 (deepspeed_trainer:226) INFO: 5epoch:train:7201-7300batch: iter_time=1.091e-04, loss_ctc=125.528, loss_att=112.897, acc=0.513, loss=116.633, grad_norm=4.507, loss_scale=1.000, learning_rate=1.889e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:50:53,637 (deepspeed_trainer:226) INFO: 5epoch:train:7301-7400batch: iter_time=1.076e-04, loss_ctc=111.340, loss_att=94.120, acc=0.530, loss=99.287, grad_norm=3.867, loss_scale=1.000, learning_rate=1.888e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 00:51:30,771 (deepspeed_trainer:226) INFO: 5epoch:train:7401-7500batch: iter_time=1.076e-04, loss_ctc=107.979, loss_att=95.461, acc=0.539, loss=99.165, grad_norm=3.876, loss_scale=1.000, learning_rate=1.886e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 00:51:33,858 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-29 00:52:00,637 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 00:52:06,298 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 00:52:06,298 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-29 00:52:06,300 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 00:52:33,704] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:26,950] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:34,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:34,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:27,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:27,842] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:34,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:34,899] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:35,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:35,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:28,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:29,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:29,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:29,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:36,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:52:29,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:22,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:22,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:22,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:22,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:16,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:16,589] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:23,567] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:16,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:23,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:17,450] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:17,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:24,409] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:17,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:24,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:17,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:53:18,078] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:10,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:11,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:04,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:12,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:05,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:05,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:05,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:12,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:13,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:06,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:13,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:06,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:07,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:13,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:07,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:14,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:58,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:58,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:52,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:59,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:53,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:53,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:55:00,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:54,550] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:55,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:55,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:55,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:55:03,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:55:03,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:55:03,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:55:03,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 00:54:57,121] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 00:56:08,607 (deepspeed_trainer:226) INFO: 5epoch:train:7501-7600batch: iter_time=2.387, loss_ctc=109.904, loss_att=91.776, acc=0.532, loss=97.214, grad_norm=3.696, loss_scale=1.000, learning_rate=1.885e-04, step_time=0.391 [cnode7-008:0/16] 2024-11-29 00:56:46,555 (deepspeed_trainer:226) INFO: 5epoch:train:7601-7700batch: iter_time=1.455e-04, loss_ctc=103.631, loss_att=91.424, acc=0.537, loss=95.058, grad_norm=3.494, loss_scale=1.000, learning_rate=1.884e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 00:57:23,952 (deepspeed_trainer:226) INFO: 5epoch:train:7701-7800batch: iter_time=1.117e-04, loss_ctc=118.665, loss_att=109.791, acc=0.534, loss=112.448, grad_norm=3.336, loss_scale=1.000, learning_rate=1.882e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 00:58:01,633 (deepspeed_trainer:226) INFO: 5epoch:train:7801-7900batch: iter_time=1.079e-04, loss_ctc=122.964, loss_att=112.051, acc=0.524, loss=115.362, grad_norm=4.050, loss_scale=1.000, learning_rate=1.881e-04, step_time=0.377 [2024-11-29 00:58:38,389] [INFO] [logging.py:129:log_dist] [Rank 0] step=68000, skipped=0, lr=[np.float64(0.00018786590596366775)], mom=[[0.9, 0.98]] [2024-11-29 00:58:38,390] [INFO] [timer.py:264:stop] epoch=0/micro_step=68000/global_step=68000, RunningAvgSamplesPerSec=42.90788529987202, CurrSamplesPerSec=46.262634382447814, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 00:58:38,391 (deepspeed_trainer:226) INFO: 5epoch:train:7901-8000batch: iter_time=1.070e-04, loss_ctc=108.388, loss_att=93.916, acc=0.517, loss=98.256, grad_norm=3.617, loss_scale=1.000, learning_rate=1.879e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 00:59:15,483 (deepspeed_trainer:226) INFO: 5epoch:train:8001-8100batch: iter_time=1.076e-04, loss_ctc=116.729, loss_att=96.592, acc=0.553, loss=102.600, grad_norm=4.064, loss_scale=1.000, learning_rate=1.878e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 00:59:52,482 (deepspeed_trainer:226) INFO: 5epoch:train:8101-8200batch: iter_time=1.100e-04, loss_ctc=120.157, loss_att=106.054, acc=0.542, loss=110.285, grad_norm=3.789, loss_scale=1.000, learning_rate=1.877e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 01:00:29,177 (deepspeed_trainer:226) INFO: 5epoch:train:8201-8300batch: iter_time=1.096e-04, loss_ctc=105.998, loss_att=92.032, acc=0.532, loss=96.243, grad_norm=3.794, loss_scale=1.000, learning_rate=1.875e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 01:01:07,405 (deepspeed_trainer:226) INFO: 5epoch:train:8301-8400batch: iter_time=1.088e-04, loss_ctc=109.683, loss_att=97.758, acc=0.539, loss=101.326, grad_norm=4.645, loss_scale=1.000, learning_rate=1.874e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 01:01:44,682 (deepspeed_trainer:226) INFO: 5epoch:train:8401-8500batch: iter_time=1.214e-04, loss_ctc=118.412, loss_att=97.692, acc=0.545, loss=103.918, grad_norm=3.943, loss_scale=1.000, learning_rate=1.872e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 01:02:21,933 (deepspeed_trainer:226) INFO: 5epoch:train:8501-8600batch: iter_time=1.089e-04, loss_ctc=109.020, loss_att=103.002, acc=0.525, loss=104.791, grad_norm=3.977, loss_scale=1.000, learning_rate=1.871e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 01:02:59,350 (deepspeed_trainer:226) INFO: 5epoch:train:8601-8700batch: iter_time=1.078e-04, loss_ctc=105.953, loss_att=90.384, acc=0.530, loss=95.054, grad_norm=4.108, loss_scale=1.000, learning_rate=1.870e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 01:03:38,622 (deepspeed_trainer:226) INFO: 5epoch:train:8701-8800batch: iter_time=1.084e-04, loss_ctc=111.662, loss_att=103.026, acc=0.538, loss=105.603, grad_norm=3.726, loss_scale=1.000, learning_rate=1.868e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 01:04:17,922 (deepspeed_trainer:226) INFO: 5epoch:train:8801-8900batch: iter_time=1.085e-04, loss_ctc=121.279, loss_att=93.781, acc=0.562, loss=102.065, grad_norm=3.722, loss_scale=1.000, learning_rate=1.867e-04, step_time=0.393 [2024-11-29 01:04:55,449] [INFO] [logging.py:129:log_dist] [Rank 0] step=69000, skipped=0, lr=[np.float64(0.0001864996102065085)], mom=[[0.9, 0.98]] [2024-11-29 01:04:55,449] [INFO] [timer.py:264:stop] epoch=0/micro_step=69000/global_step=69000, RunningAvgSamplesPerSec=42.91304703958343, CurrSamplesPerSec=43.55325291263939, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:04:55,451 (deepspeed_trainer:226) INFO: 5epoch:train:8901-9000batch: iter_time=1.081e-04, loss_ctc=119.257, loss_att=106.912, acc=0.528, loss=110.634, grad_norm=4.473, loss_scale=1.000, learning_rate=1.866e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:05:32,555 (deepspeed_trainer:226) INFO: 5epoch:train:9001-9100batch: iter_time=1.109e-04, loss_ctc=115.030, loss_att=98.421, acc=0.541, loss=103.424, grad_norm=4.457, loss_scale=1.000, learning_rate=1.864e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 01:06:10,041 (deepspeed_trainer:226) INFO: 5epoch:train:9101-9200batch: iter_time=1.074e-04, loss_ctc=113.549, loss_att=100.688, acc=0.529, loss=104.516, grad_norm=3.872, loss_scale=1.000, learning_rate=1.863e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:06:47,520 (deepspeed_trainer:226) INFO: 5epoch:train:9201-9300batch: iter_time=1.088e-04, loss_ctc=111.918, loss_att=94.901, acc=0.531, loss=100.017, grad_norm=3.413, loss_scale=1.000, learning_rate=1.862e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 01:07:19,207 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-29 01:07:45,120 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 01:07:50,565 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 01:07:50,565 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-29 01:07:50,567 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 01:08:11,450] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:18,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:18,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:18,839] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:19,033] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:12,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:19,688] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:12,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:12,975] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:13,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:20,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:13,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:13,449] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:13,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:20,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:08:22,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:07,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:07,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:07,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:00,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:08,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:08,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:01,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:08,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:01,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:01,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:09,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:02,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:02,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:02,804] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:10,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:03,120] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:55,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:49,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:56,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:56,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:57,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:50,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:50,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:57,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:57,815] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:50,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:58,101] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:51,307] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:51,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:58,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:51,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:09:52,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:37,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:44,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:45,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:45,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:39,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:46,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:39,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:46,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:46,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:47,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:40,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:40,151] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:47,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:40,899] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:40,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:10:40,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 01:11:24,737 (deepspeed_trainer:226) INFO: 5epoch:train:9301-9400batch: iter_time=2.373, loss_ctc=108.826, loss_att=92.317, acc=0.551, loss=97.297, grad_norm=3.437, loss_scale=1.000, learning_rate=1.860e-04, step_time=0.399 [cnode7-008:0/16] 2024-11-29 01:12:02,154 (deepspeed_trainer:226) INFO: 5epoch:train:9401-9500batch: iter_time=1.021e-04, loss_ctc=102.699, loss_att=87.633, acc=0.545, loss=92.109, grad_norm=3.288, loss_scale=1.000, learning_rate=1.859e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 01:12:39,514 (deepspeed_trainer:226) INFO: 5epoch:train:9501-9600batch: iter_time=1.053e-04, loss_ctc=96.638, loss_att=87.644, acc=0.545, loss=90.339, grad_norm=3.407, loss_scale=1.000, learning_rate=1.858e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 01:13:17,918 (deepspeed_trainer:226) INFO: 5epoch:train:9601-9700batch: iter_time=1.089e-04, loss_ctc=132.649, loss_att=116.480, acc=0.550, loss=121.323, grad_norm=3.773, loss_scale=1.000, learning_rate=1.856e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 01:13:57,202 (deepspeed_trainer:226) INFO: 5epoch:train:9701-9800batch: iter_time=1.070e-04, loss_ctc=106.195, loss_att=95.271, acc=0.531, loss=98.555, grad_norm=3.735, loss_scale=1.000, learning_rate=1.855e-04, step_time=0.393 [cnode7-008:0/16] 2024-11-29 01:14:36,033 (deepspeed_trainer:226) INFO: 5epoch:train:9801-9900batch: iter_time=1.082e-04, loss_ctc=106.480, loss_att=94.168, acc=0.530, loss=97.883, grad_norm=3.504, loss_scale=1.000, learning_rate=1.854e-04, step_time=0.388 [2024-11-29 01:15:13,689] [INFO] [logging.py:129:log_dist] [Rank 0] step=70000, skipped=0, lr=[np.float64(0.00018516269736853832)], mom=[[0.9, 0.98]] [2024-11-29 01:15:13,690] [INFO] [timer.py:264:stop] epoch=0/micro_step=70000/global_step=70000, RunningAvgSamplesPerSec=42.91147068334334, CurrSamplesPerSec=47.031884153899966, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:15:13,691 (deepspeed_trainer:226) INFO: 5epoch:train:9901-10000batch: iter_time=1.119e-04, loss_ctc=118.168, loss_att=92.351, acc=0.573, loss=100.109, grad_norm=4.103, loss_scale=1.000, learning_rate=1.852e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:15:51,440 (deepspeed_trainer:226) INFO: 5epoch:train:10001-10100batch: iter_time=1.083e-04, loss_ctc=131.454, loss_att=118.945, acc=0.535, loss=122.659, grad_norm=4.240, loss_scale=1.000, learning_rate=1.851e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 01:16:28,271 (deepspeed_trainer:226) INFO: 5epoch:train:10101-10200batch: iter_time=1.131e-04, loss_ctc=91.351, loss_att=80.925, acc=0.547, loss=84.036, grad_norm=3.401, loss_scale=1.000, learning_rate=1.850e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 01:17:05,026 (deepspeed_trainer:226) INFO: 5epoch:train:10201-10300batch: iter_time=1.069e-04, loss_ctc=105.078, loss_att=86.132, acc=0.557, loss=91.832, grad_norm=3.567, loss_scale=1.000, learning_rate=1.848e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 01:17:42,913 (deepspeed_trainer:226) INFO: 5epoch:train:10301-10400batch: iter_time=1.068e-04, loss_ctc=119.574, loss_att=102.000, acc=0.550, loss=107.270, grad_norm=4.083, loss_scale=1.000, learning_rate=1.847e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 01:18:19,917 (deepspeed_trainer:226) INFO: 5epoch:train:10401-10500batch: iter_time=1.064e-04, loss_ctc=103.925, loss_att=92.105, acc=0.541, loss=95.636, grad_norm=3.853, loss_scale=1.000, learning_rate=1.846e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 01:18:56,887 (deepspeed_trainer:226) INFO: 5epoch:train:10501-10600batch: iter_time=1.076e-04, loss_ctc=105.873, loss_att=91.804, acc=0.540, loss=95.998, grad_norm=4.621, loss_scale=1.000, learning_rate=1.844e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 01:19:34,390 (deepspeed_trainer:226) INFO: 5epoch:train:10601-10700batch: iter_time=1.078e-04, loss_ctc=117.367, loss_att=103.204, acc=0.550, loss=107.415, grad_norm=3.831, loss_scale=1.000, learning_rate=1.843e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 01:20:12,292 (deepspeed_trainer:226) INFO: 5epoch:train:10701-10800batch: iter_time=1.082e-04, loss_ctc=107.834, loss_att=82.029, acc=0.571, loss=89.757, grad_norm=3.321, loss_scale=1.000, learning_rate=1.842e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 01:20:50,236 (deepspeed_trainer:226) INFO: 5epoch:train:10801-10900batch: iter_time=1.076e-04, loss_ctc=122.698, loss_att=110.356, acc=0.537, loss=114.071, grad_norm=3.829, loss_scale=1.000, learning_rate=1.840e-04, step_time=0.379 [2024-11-29 01:21:27,393] [INFO] [logging.py:129:log_dist] [Rank 0] step=71000, skipped=0, lr=[np.float64(0.00018385412916793913)], mom=[[0.9, 0.98]] [2024-11-29 01:21:27,393] [INFO] [timer.py:264:stop] epoch=0/micro_step=71000/global_step=71000, RunningAvgSamplesPerSec=42.9218595814493, CurrSamplesPerSec=46.20988156492091, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:21:27,395 (deepspeed_trainer:226) INFO: 5epoch:train:10901-11000batch: iter_time=1.089e-04, loss_ctc=111.522, loss_att=99.294, acc=0.555, loss=102.986, grad_norm=3.511, loss_scale=1.000, learning_rate=1.839e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 01:22:05,335 (deepspeed_trainer:226) INFO: 5epoch:train:11001-11100batch: iter_time=1.063e-04, loss_ctc=108.679, loss_att=91.248, acc=0.534, loss=96.461, grad_norm=3.658, loss_scale=1.000, learning_rate=1.838e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 01:22:43,808 (deepspeed_trainer:226) INFO: 5epoch:train:11101-11200batch: iter_time=1.067e-04, loss_ctc=104.302, loss_att=87.188, acc=0.541, loss=92.324, grad_norm=3.866, loss_scale=1.000, learning_rate=1.837e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 01:23:06,629 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-29 01:23:33,820 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 01:23:39,317 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 01:23:39,317 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-29 01:23:39,319 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 01:24:06,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:06,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:23:59,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:07,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:01,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:08,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:08,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:08,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:09,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:09,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:01,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:02,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:02,120] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:02,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:02,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:02,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:54,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:55,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:47,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:56,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:49,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:49,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:56,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:49,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:57,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:57,625] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:57,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:50,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:58,066] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:50,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:51,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:24:51,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:42,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:35,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:43,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:36,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:45,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:45,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:37,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:45,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:38,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:38,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:46,307] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:39,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:46,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:39,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:46,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:25:39,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:22,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:31,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:32,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:24,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:25,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:26,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:26,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:34,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:26,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:27,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:34,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:35,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:35,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:35,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:35,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:26:30,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 01:27:22,583 (deepspeed_trainer:226) INFO: 5epoch:train:11201-11300batch: iter_time=2.348, loss_ctc=113.818, loss_att=100.960, acc=0.544, loss=104.845, grad_norm=3.880, loss_scale=1.000, learning_rate=1.835e-04, step_time=0.440 [cnode7-008:0/16] 2024-11-29 01:28:00,249 (deepspeed_trainer:226) INFO: 5epoch:train:11301-11400batch: iter_time=1.020e-04, loss_ctc=102.737, loss_att=86.233, acc=0.553, loss=91.200, grad_norm=3.639, loss_scale=1.000, learning_rate=1.834e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:28:37,694 (deepspeed_trainer:226) INFO: 5epoch:train:11401-11500batch: iter_time=1.039e-04, loss_ctc=98.170, loss_att=90.803, acc=0.548, loss=93.066, grad_norm=3.511, loss_scale=1.000, learning_rate=1.833e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 01:29:15,028 (deepspeed_trainer:226) INFO: 5epoch:train:11501-11600batch: iter_time=1.070e-04, loss_ctc=129.924, loss_att=112.898, acc=0.546, loss=118.032, grad_norm=4.702, loss_scale=1.000, learning_rate=1.831e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 01:29:51,858 (deepspeed_trainer:226) INFO: 5epoch:train:11601-11700batch: iter_time=1.060e-04, loss_ctc=96.256, loss_att=88.694, acc=0.529, loss=90.936, grad_norm=4.613, loss_scale=1.000, learning_rate=1.830e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 01:30:29,452 (deepspeed_trainer:226) INFO: 5epoch:train:11701-11800batch: iter_time=1.067e-04, loss_ctc=120.693, loss_att=101.891, acc=0.543, loss=107.590, grad_norm=4.304, loss_scale=1.000, learning_rate=1.829e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:31:07,102 (deepspeed_trainer:226) INFO: 5epoch:train:11801-11900batch: iter_time=1.080e-04, loss_ctc=111.999, loss_att=92.767, acc=0.557, loss=98.518, grad_norm=3.561, loss_scale=1.000, learning_rate=1.828e-04, step_time=0.376 [2024-11-29 01:31:43,614] [INFO] [logging.py:129:log_dist] [Rank 0] step=72000, skipped=0, lr=[np.float64(0.00018257291797197175)], mom=[[0.9, 0.98]] [2024-11-29 01:31:43,615] [INFO] [timer.py:264:stop] epoch=0/micro_step=72000/global_step=72000, RunningAvgSamplesPerSec=42.91944486977935, CurrSamplesPerSec=45.857455492849965, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:31:43,616 (deepspeed_trainer:226) INFO: 5epoch:train:11901-12000batch: iter_time=1.059e-04, loss_ctc=111.869, loss_att=99.281, acc=0.548, loss=103.047, grad_norm=3.698, loss_scale=1.000, learning_rate=1.826e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 01:32:20,026 (deepspeed_trainer:226) INFO: 5epoch:train:12001-12100batch: iter_time=1.072e-04, loss_ctc=96.509, loss_att=87.424, acc=0.545, loss=90.162, grad_norm=3.840, loss_scale=1.000, learning_rate=1.825e-04, step_time=0.364 [cnode7-008:0/16] 2024-11-29 01:32:58,821 (deepspeed_trainer:226) INFO: 5epoch:train:12101-12200batch: iter_time=1.070e-04, loss_ctc=112.694, loss_att=90.109, acc=0.549, loss=96.851, grad_norm=4.301, loss_scale=1.000, learning_rate=1.824e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 01:33:36,334 (deepspeed_trainer:226) INFO: 5epoch:train:12201-12300batch: iter_time=1.090e-04, loss_ctc=100.834, loss_att=87.323, acc=0.553, loss=91.347, grad_norm=3.590, loss_scale=1.000, learning_rate=1.823e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:34:15,048 (deepspeed_trainer:226) INFO: 5epoch:train:12301-12400batch: iter_time=1.083e-04, loss_ctc=103.879, loss_att=90.863, acc=0.543, loss=94.787, grad_norm=4.018, loss_scale=1.000, learning_rate=1.821e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 01:34:52,757 (deepspeed_trainer:226) INFO: 5epoch:train:12401-12500batch: iter_time=1.060e-04, loss_ctc=106.820, loss_att=98.441, acc=0.536, loss=100.959, grad_norm=3.814, loss_scale=1.000, learning_rate=1.820e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:35:30,488 (deepspeed_trainer:226) INFO: 5epoch:train:12501-12600batch: iter_time=1.153e-04, loss_ctc=119.189, loss_att=102.410, acc=0.555, loss=107.466, grad_norm=3.466, loss_scale=1.000, learning_rate=1.819e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 01:36:07,961 (deepspeed_trainer:226) INFO: 5epoch:train:12601-12700batch: iter_time=1.081e-04, loss_ctc=105.204, loss_att=80.086, acc=0.568, loss=87.654, grad_norm=3.592, loss_scale=1.000, learning_rate=1.818e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:36:45,480 (deepspeed_trainer:226) INFO: 5epoch:train:12701-12800batch: iter_time=1.095e-04, loss_ctc=120.713, loss_att=107.676, acc=0.536, loss=111.571, grad_norm=4.179, loss_scale=1.000, learning_rate=1.816e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:37:22,612 (deepspeed_trainer:226) INFO: 5epoch:train:12801-12900batch: iter_time=1.078e-04, loss_ctc=111.388, loss_att=99.394, acc=0.546, loss=102.976, grad_norm=4.501, loss_scale=1.000, learning_rate=1.815e-04, step_time=0.371 [2024-11-29 01:38:00,100] [INFO] [logging.py:129:log_dist] [Rank 0] step=73000, skipped=0, lr=[np.float64(0.0001813181236639463)], mom=[[0.9, 0.98]] [2024-11-29 01:38:00,101] [INFO] [timer.py:264:stop] epoch=0/micro_step=73000/global_step=73000, RunningAvgSamplesPerSec=42.92490571217306, CurrSamplesPerSec=46.10388083123847, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:38:00,102 (deepspeed_trainer:226) INFO: 5epoch:train:12901-13000batch: iter_time=1.035e-04, loss_ctc=107.039, loss_att=88.001, acc=0.548, loss=93.695, grad_norm=3.625, loss_scale=1.000, learning_rate=1.814e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 01:38:37,253 (deepspeed_trainer:226) INFO: 5epoch:train:13001-13100batch: iter_time=1.029e-04, loss_ctc=105.475, loss_att=89.312, acc=0.545, loss=94.142, grad_norm=3.893, loss_scale=1.000, learning_rate=1.813e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 01:38:49,410 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-29 01:39:16,486 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 01:39:21,987 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 01:39:21,988 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-29 01:39:21,990 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 01:39:49,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:42,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:50,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:42,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:51,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:43,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:44,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:51,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:51,820] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:44,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:51,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:44,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:44,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:44,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:52,033] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:39:53,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:37,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:30,604] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:38,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:31,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:31,782] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:39,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:32,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:40,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:40,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:33,289] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:33,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:33,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:41,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:41,237] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:33,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:40:45,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:19,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:26,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:19,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:27,483] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:20,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:20,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:21,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:29,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:29,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:30,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:30,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:30,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:23,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:23,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:26,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:41:37,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:06,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:07,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:16,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:16,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:08,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:09,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:18,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:18,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:11,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:18,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:19,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:19,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:11,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:12,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:16,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:42:28,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 01:43:22,085 (deepspeed_trainer:226) INFO: 5epoch:train:13101-13200batch: iter_time=2.472, loss_ctc=103.073, loss_att=85.593, acc=0.555, loss=90.816, grad_norm=3.556, loss_scale=1.000, learning_rate=1.811e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:44:00,920 (deepspeed_trainer:226) INFO: 5epoch:train:13201-13300batch: iter_time=1.095e-04, loss_ctc=99.182, loss_att=82.779, acc=0.560, loss=87.662, grad_norm=3.343, loss_scale=1.000, learning_rate=1.810e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 01:44:38,953 (deepspeed_trainer:226) INFO: 5epoch:train:13301-13400batch: iter_time=1.067e-04, loss_ctc=108.567, loss_att=97.793, acc=0.559, loss=101.044, grad_norm=3.499, loss_scale=1.000, learning_rate=1.809e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:45:16,275 (deepspeed_trainer:226) INFO: 5epoch:train:13401-13500batch: iter_time=1.082e-04, loss_ctc=119.881, loss_att=102.133, acc=0.556, loss=107.419, grad_norm=3.844, loss_scale=1.000, learning_rate=1.808e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 01:45:53,334 (deepspeed_trainer:226) INFO: 5epoch:train:13501-13600batch: iter_time=1.062e-04, loss_ctc=98.112, loss_att=83.111, acc=0.539, loss=87.614, grad_norm=3.476, loss_scale=1.000, learning_rate=1.806e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 01:46:30,507 (deepspeed_trainer:226) INFO: 5epoch:train:13601-13700batch: iter_time=1.103e-04, loss_ctc=113.909, loss_att=97.369, acc=0.553, loss=102.391, grad_norm=3.886, loss_scale=1.000, learning_rate=1.805e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 01:47:07,465 (deepspeed_trainer:226) INFO: 5epoch:train:13701-13800batch: iter_time=1.125e-04, loss_ctc=114.082, loss_att=90.717, acc=0.576, loss=97.723, grad_norm=3.326, loss_scale=1.000, learning_rate=1.804e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 01:47:44,201 (deepspeed_trainer:226) INFO: 5epoch:train:13801-13900batch: iter_time=1.132e-04, loss_ctc=105.041, loss_att=94.013, acc=0.558, loss=97.368, grad_norm=3.403, loss_scale=1.000, learning_rate=1.803e-04, step_time=0.367 [2024-11-29 01:48:21,019] [INFO] [logging.py:129:log_dist] [Rank 0] step=74000, skipped=0, lr=[np.float64(0.00018008885074385103)], mom=[[0.9, 0.98]] [2024-11-29 01:48:21,019] [INFO] [timer.py:264:stop] epoch=0/micro_step=74000/global_step=74000, RunningAvgSamplesPerSec=42.93453560151083, CurrSamplesPerSec=44.78692000598124, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:48:21,020 (deepspeed_trainer:226) INFO: 5epoch:train:13901-14000batch: iter_time=1.129e-04, loss_ctc=97.289, loss_att=85.558, acc=0.552, loss=89.064, grad_norm=3.465, loss_scale=1.000, learning_rate=1.801e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 01:48:58,031 (deepspeed_trainer:226) INFO: 5epoch:train:14001-14100batch: iter_time=1.081e-04, loss_ctc=110.198, loss_att=89.894, acc=0.559, loss=96.016, grad_norm=3.652, loss_scale=1.000, learning_rate=1.800e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 01:49:34,568 (deepspeed_trainer:226) INFO: 5epoch:train:14101-14200batch: iter_time=1.057e-04, loss_ctc=103.987, loss_att=90.606, acc=0.557, loss=94.648, grad_norm=3.780, loss_scale=1.000, learning_rate=1.799e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 01:50:10,933 (deepspeed_trainer:226) INFO: 5epoch:train:14201-14300batch: iter_time=1.113e-04, loss_ctc=100.249, loss_att=81.346, acc=0.552, loss=87.002, grad_norm=4.140, loss_scale=1.000, learning_rate=1.798e-04, step_time=0.363 [cnode7-008:0/16] 2024-11-29 01:50:48,550 (deepspeed_trainer:226) INFO: 5epoch:train:14301-14400batch: iter_time=1.167e-04, loss_ctc=106.955, loss_att=100.244, acc=0.544, loss=102.313, grad_norm=3.567, loss_scale=1.000, learning_rate=1.797e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 01:51:25,886 (deepspeed_trainer:226) INFO: 5epoch:train:14401-14500batch: iter_time=1.184e-04, loss_ctc=118.746, loss_att=93.736, acc=0.572, loss=101.257, grad_norm=3.360, loss_scale=1.000, learning_rate=1.795e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 01:52:02,495 (deepspeed_trainer:226) INFO: 5epoch:train:14501-14600batch: iter_time=1.155e-04, loss_ctc=106.265, loss_att=85.141, acc=0.566, loss=91.493, grad_norm=3.455, loss_scale=1.000, learning_rate=1.794e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 01:52:41,090 (deepspeed_trainer:226) INFO: 5epoch:train:14601-14700batch: iter_time=1.178e-04, loss_ctc=109.799, loss_att=98.472, acc=0.549, loss=101.872, grad_norm=3.692, loss_scale=1.000, learning_rate=1.793e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 01:53:19,198 (deepspeed_trainer:226) INFO: 5epoch:train:14701-14800batch: iter_time=1.134e-04, loss_ctc=115.959, loss_att=97.906, acc=0.547, loss=103.343, grad_norm=4.198, loss_scale=1.000, learning_rate=1.792e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 01:53:57,959 (deepspeed_trainer:226) INFO: 5epoch:train:14801-14900batch: iter_time=1.117e-04, loss_ctc=103.528, loss_att=84.251, acc=0.559, loss=90.025, grad_norm=3.189, loss_scale=1.000, learning_rate=1.791e-04, step_time=0.387 [2024-11-29 01:54:34,981] [INFO] [logging.py:129:log_dist] [Rank 0] step=75000, skipped=0, lr=[np.float64(0.00017888424564232075)], mom=[[0.9, 0.98]] [2024-11-29 01:54:34,981] [INFO] [timer.py:264:stop] epoch=0/micro_step=75000/global_step=75000, RunningAvgSamplesPerSec=42.94346220433027, CurrSamplesPerSec=44.157070134088606, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 01:54:34,982 (deepspeed_trainer:226) INFO: 5epoch:train:14901-15000batch: iter_time=1.068e-04, loss_ctc=100.773, loss_att=84.628, acc=0.568, loss=89.463, grad_norm=3.299, loss_scale=1.000, learning_rate=1.789e-04, step_time=0.370 [2024-11-29 01:54:50,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:42,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:42,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,249] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,289] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:43,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:51,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:58,314] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:58,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:06,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:58,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:06,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:58,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:07,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:59,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:54:59,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:00,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:00,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:08,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:08,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:08,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:08,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:08,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:13,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:14,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:14,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:14,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:15,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:23,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:23,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:15,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:15,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:16,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:24,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:24,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:24,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:25,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:25,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:25,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:29,253] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:31,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:31,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:31,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:31,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:39,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:39,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:32,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:32,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:40,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:32,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:40,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:40,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:41,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:41,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:41,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:55:53,873] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 5 is about to be saved! [2024-11-29 01:55:53,904] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/mp_rank_00_model_states.pt [2024-11-29 01:55:53,904] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/mp_rank_00_model_states.pt... [2024-11-29 01:55:55,973] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/mp_rank_00_model_states.pt. [2024-11-29 01:55:48,292] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,132] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,132] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,133] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,295] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,136] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,298] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,136] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,298] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,138] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,138] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,300] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,300] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,300] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,301] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-29 01:55:56,140] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-29 01:55:48,798] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-29 01:55:48,798] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-29 01:55:48,798] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,725] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,729] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,729] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,739] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,739] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,739] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:48,929] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-29 01:55:48,930] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-29 01:55:48,930] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:48,939] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-29 01:55:48,939] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-29 01:55:48,939] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:48,996] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-29 01:55:48,996] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-29 01:55:48,996] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,857] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,857] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,857] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,872] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,872] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,872] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,888] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,888] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,888] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,898] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,898] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,898] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:49,065] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-29 01:55:49,065] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-29 01:55:49,065] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,906] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,907] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,907] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:56,913] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-29 01:55:56,914] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-29 01:55:56,914] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:49,095] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-29 01:55:49,095] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-29 01:55:49,095] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:49,140] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-29 01:55:49,140] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-29 01:55:49,141] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [2024-11-29 01:55:49,167] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-29 01:55:49,167] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_5/5/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-29 01:55:49,168] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 5 is ready now! [cnode7-008:0/16] 2024-11-29 01:55:57,015 (deepspeed_trainer:158) INFO: 5epoch results: [train] iter_time=0.127, loss_ctc=115.696, loss_att=101.563, acc=0.523, loss=105.805, grad_norm=3.840, loss_scale=1.000, learning_rate=1.889e-04, step_time=0.377, time=2 hours, 6 minutes and 13.4 seconds, total_count=75005, gpu_max_cached_mem_GB=27.471, [valid] loss_ctc=15.312, cer_ctc=0.417, loss_att=16.875, acc=0.685, cer=0.516, wer=1.000, loss=16.375, time=1 minute and 10.02 seconds, total_count=5, gpu_max_cached_mem_GB=27.471 [cnode7-008:0/16] 2024-11-29 01:55:58,641 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-29 01:56:25,940 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 01:56:31,672 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 01:56:31,672 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-29 01:56:31,674 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 01:56:44,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:45,473] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:53,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:45,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:53,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:46,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:46,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:46,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:46,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:54,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:47,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:55,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:55,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:56,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:56,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:56:57,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:32,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:40,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:33,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:33,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:33,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:33,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:34,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:42,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:34,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:42,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:35,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:43,325] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:43,686] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:43,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:44,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:57:45,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:19,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:20,236] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:20,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:20,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:21,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:29,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:21,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:29,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:30,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:30,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:23,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:23,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:31,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:31,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:32,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:58:33,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:06,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:07,289] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:08,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:08,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:16,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:09,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:09,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:17,743] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:10,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:18,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:19,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:11,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:19,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:19,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:20,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 01:59:21,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 02:00:26,659 (deepspeed_trainer:226) INFO: 6epoch:train:1-100batch: iter_time=2.259, loss_ctc=101.285, loss_att=87.985, acc=0.557, loss=91.984, grad_norm=3.621, loss_scale=1.000, learning_rate=1.788e-04, step_time=0.421 [cnode7-008:0/16] 2024-11-29 02:01:03,910 (deepspeed_trainer:226) INFO: 6epoch:train:101-200batch: iter_time=1.049e-04, loss_ctc=115.537, loss_att=94.305, acc=0.543, loss=100.667, grad_norm=3.662, loss_scale=1.000, learning_rate=1.787e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:01:41,204 (deepspeed_trainer:226) INFO: 6epoch:train:201-300batch: iter_time=1.073e-04, loss_ctc=100.841, loss_att=91.329, acc=0.559, loss=94.138, grad_norm=3.352, loss_scale=1.000, learning_rate=1.786e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:02:18,645 (deepspeed_trainer:226) INFO: 6epoch:train:301-400batch: iter_time=1.088e-04, loss_ctc=102.296, loss_att=85.004, acc=0.568, loss=90.170, grad_norm=3.132, loss_scale=1.000, learning_rate=1.785e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:02:56,654 (deepspeed_trainer:226) INFO: 6epoch:train:401-500batch: iter_time=1.094e-04, loss_ctc=111.122, loss_att=104.772, acc=0.557, loss=106.670, grad_norm=3.455, loss_scale=1.000, learning_rate=1.783e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:03:33,869 (deepspeed_trainer:226) INFO: 6epoch:train:501-600batch: iter_time=1.071e-04, loss_ctc=125.474, loss_att=106.665, acc=0.547, loss=112.339, grad_norm=4.831, loss_scale=1.000, learning_rate=1.782e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:04:11,750 (deepspeed_trainer:226) INFO: 6epoch:train:601-700batch: iter_time=1.109e-04, loss_ctc=108.407, loss_att=100.938, acc=0.545, loss=103.198, grad_norm=4.410, loss_scale=1.000, learning_rate=1.781e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:04:51,019 (deepspeed_trainer:226) INFO: 6epoch:train:701-800batch: iter_time=1.086e-04, loss_ctc=113.753, loss_att=101.201, acc=0.553, loss=104.978, grad_norm=3.766, loss_scale=1.000, learning_rate=1.780e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 02:05:29,047 (deepspeed_trainer:226) INFO: 6epoch:train:801-900batch: iter_time=1.085e-04, loss_ctc=102.533, loss_att=92.614, acc=0.569, loss=95.601, grad_norm=3.342, loss_scale=1.000, learning_rate=1.779e-04, step_time=0.380 [2024-11-29 02:06:06,958] [INFO] [logging.py:129:log_dist] [Rank 0] step=76000, skipped=0, lr=[np.float64(0.0001777034942296377)], mom=[[0.9, 0.98]] [2024-11-29 02:06:06,959] [INFO] [timer.py:264:stop] epoch=0/micro_step=76000/global_step=76000, RunningAvgSamplesPerSec=42.93924689618517, CurrSamplesPerSec=45.789213937033864, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:06:06,960 (deepspeed_trainer:226) INFO: 6epoch:train:901-1000batch: iter_time=1.067e-04, loss_ctc=86.959, loss_att=76.858, acc=0.568, loss=79.885, grad_norm=3.177, loss_scale=1.000, learning_rate=1.778e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:06:44,184 (deepspeed_trainer:226) INFO: 6epoch:train:1001-1100batch: iter_time=1.069e-04, loss_ctc=105.319, loss_att=84.559, acc=0.575, loss=90.817, grad_norm=3.583, loss_scale=1.000, learning_rate=1.776e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:07:21,200 (deepspeed_trainer:226) INFO: 6epoch:train:1101-1200batch: iter_time=1.062e-04, loss_ctc=97.708, loss_att=79.258, acc=0.575, loss=84.816, grad_norm=3.894, loss_scale=1.000, learning_rate=1.775e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 02:07:58,380 (deepspeed_trainer:226) INFO: 6epoch:train:1201-1300batch: iter_time=1.061e-04, loss_ctc=101.490, loss_att=84.251, acc=0.578, loss=89.469, grad_norm=3.540, loss_scale=1.000, learning_rate=1.774e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:08:35,828 (deepspeed_trainer:226) INFO: 6epoch:train:1301-1400batch: iter_time=1.105e-04, loss_ctc=97.782, loss_att=90.689, acc=0.559, loss=92.825, grad_norm=3.444, loss_scale=1.000, learning_rate=1.773e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:09:13,274 (deepspeed_trainer:226) INFO: 6epoch:train:1401-1500batch: iter_time=1.074e-04, loss_ctc=116.295, loss_att=94.666, acc=0.579, loss=101.161, grad_norm=3.841, loss_scale=1.000, learning_rate=1.772e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:09:50,859 (deepspeed_trainer:226) INFO: 6epoch:train:1501-1600batch: iter_time=1.067e-04, loss_ctc=107.363, loss_att=98.993, acc=0.556, loss=101.526, grad_norm=3.826, loss_scale=1.000, learning_rate=1.771e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 02:10:28,224 (deepspeed_trainer:226) INFO: 6epoch:train:1601-1700batch: iter_time=1.075e-04, loss_ctc=112.230, loss_att=91.076, acc=0.563, loss=97.396, grad_norm=4.122, loss_scale=1.000, learning_rate=1.769e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 02:11:05,659 (deepspeed_trainer:226) INFO: 6epoch:train:1701-1800batch: iter_time=1.066e-04, loss_ctc=111.070, loss_att=100.576, acc=0.552, loss=103.737, grad_norm=4.144, loss_scale=1.000, learning_rate=1.768e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:11:38,308 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-29 02:12:06,139 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 02:12:11,798 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 02:12:11,798 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-29 02:12:11,831 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 02:12:34,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,156] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:35,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:44,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:12:45,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:26,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:28,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:36,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:36,737] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:37,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:37,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:30,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:38,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:38,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:38,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:13:38,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:15,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:16,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:17,173] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:17,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:17,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:18,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:18,101] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:25,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:25,512] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:26,000] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:27,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:27,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:21,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:28,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:28,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:14:28,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:02,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:05,053] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:05,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:05,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:05,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:06,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:06,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:13,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:15,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:15,655] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:16,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:17,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:17,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:18,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:11,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:15:18,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 02:15:55,667 (deepspeed_trainer:226) INFO: 6epoch:train:1801-1900batch: iter_time=2.524, loss_ctc=95.570, loss_att=77.248, acc=0.583, loss=82.725, grad_norm=3.629, loss_scale=1.000, learning_rate=1.767e-04, step_time=0.372 [2024-11-29 02:16:33,195] [INFO] [logging.py:129:log_dist] [Rank 0] step=77000, skipped=0, lr=[np.float64(0.00017654581950324663)], mom=[[0.9, 0.98]] [2024-11-29 02:16:33,198] [INFO] [timer.py:264:stop] epoch=0/micro_step=77000/global_step=77000, RunningAvgSamplesPerSec=42.948203131771244, CurrSamplesPerSec=45.05906608093798, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:16:33,200 (deepspeed_trainer:226) INFO: 6epoch:train:1901-2000batch: iter_time=1.199e-04, loss_ctc=107.422, loss_att=95.233, acc=0.549, loss=98.861, grad_norm=3.717, loss_scale=1.000, learning_rate=1.766e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:17:10,460 (deepspeed_trainer:226) INFO: 6epoch:train:2001-2100batch: iter_time=1.120e-04, loss_ctc=107.734, loss_att=89.558, acc=0.556, loss=95.007, grad_norm=3.402, loss_scale=1.000, learning_rate=1.765e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:17:47,538 (deepspeed_trainer:226) INFO: 6epoch:train:2101-2200batch: iter_time=1.118e-04, loss_ctc=97.437, loss_att=84.861, acc=0.569, loss=88.686, grad_norm=3.434, loss_scale=1.000, learning_rate=1.764e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:18:24,848 (deepspeed_trainer:226) INFO: 6epoch:train:2201-2300batch: iter_time=1.062e-04, loss_ctc=100.153, loss_att=84.877, acc=0.573, loss=89.444, grad_norm=3.059, loss_scale=1.000, learning_rate=1.763e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 02:19:02,062 (deepspeed_trainer:226) INFO: 6epoch:train:2301-2400batch: iter_time=1.051e-04, loss_ctc=114.737, loss_att=98.345, acc=0.560, loss=103.271, grad_norm=3.816, loss_scale=1.000, learning_rate=1.761e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:19:39,898 (deepspeed_trainer:226) INFO: 6epoch:train:2401-2500batch: iter_time=1.090e-04, loss_ctc=114.978, loss_att=100.165, acc=0.552, loss=104.643, grad_norm=3.765, loss_scale=1.000, learning_rate=1.760e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 02:20:17,731 (deepspeed_trainer:226) INFO: 6epoch:train:2501-2600batch: iter_time=1.075e-04, loss_ctc=108.374, loss_att=101.006, acc=0.545, loss=103.228, grad_norm=3.869, loss_scale=1.000, learning_rate=1.759e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 02:20:55,085 (deepspeed_trainer:226) INFO: 6epoch:train:2601-2700batch: iter_time=1.064e-04, loss_ctc=112.775, loss_att=102.994, acc=0.553, loss=105.953, grad_norm=3.573, loss_scale=1.000, learning_rate=1.758e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 02:21:31,885 (deepspeed_trainer:226) INFO: 6epoch:train:2701-2800batch: iter_time=1.067e-04, loss_ctc=96.176, loss_att=88.483, acc=0.563, loss=90.812, grad_norm=3.312, loss_scale=1.000, learning_rate=1.757e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 02:22:09,443 (deepspeed_trainer:226) INFO: 6epoch:train:2801-2900batch: iter_time=1.087e-04, loss_ctc=91.128, loss_att=74.079, acc=0.575, loss=79.194, grad_norm=3.461, loss_scale=1.000, learning_rate=1.756e-04, step_time=0.375 [2024-11-29 02:22:46,977] [INFO] [logging.py:129:log_dist] [Rank 0] step=78000, skipped=0, lr=[np.float64(0.00017541047943885952)], mom=[[0.9, 0.98]] [2024-11-29 02:22:46,978] [INFO] [timer.py:264:stop] epoch=0/micro_step=78000/global_step=78000, RunningAvgSamplesPerSec=42.957159615067965, CurrSamplesPerSec=42.64358903243604, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:22:46,979 (deepspeed_trainer:226) INFO: 6epoch:train:2901-3000batch: iter_time=1.074e-04, loss_ctc=97.591, loss_att=79.428, acc=0.579, loss=84.871, grad_norm=3.325, loss_scale=1.000, learning_rate=1.755e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 02:23:24,591 (deepspeed_trainer:226) INFO: 6epoch:train:3001-3100batch: iter_time=1.070e-04, loss_ctc=98.793, loss_att=77.007, acc=0.574, loss=83.562, grad_norm=3.628, loss_scale=1.000, learning_rate=1.754e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 02:24:04,202 (deepspeed_trainer:226) INFO: 6epoch:train:3101-3200batch: iter_time=1.076e-04, loss_ctc=100.369, loss_att=83.706, acc=0.583, loss=88.709, grad_norm=3.251, loss_scale=1.000, learning_rate=1.752e-04, step_time=0.396 [cnode7-008:0/16] 2024-11-29 02:24:42,351 (deepspeed_trainer:226) INFO: 6epoch:train:3201-3300batch: iter_time=1.085e-04, loss_ctc=91.521, loss_att=81.773, acc=0.576, loss=84.700, grad_norm=3.121, loss_scale=1.000, learning_rate=1.751e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 02:25:20,117 (deepspeed_trainer:226) INFO: 6epoch:train:3301-3400batch: iter_time=1.079e-04, loss_ctc=118.714, loss_att=96.894, acc=0.564, loss=103.469, grad_norm=4.151, loss_scale=1.000, learning_rate=1.750e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 02:25:57,763 (deepspeed_trainer:226) INFO: 6epoch:train:3401-3500batch: iter_time=1.081e-04, loss_ctc=100.357, loss_att=90.897, acc=0.563, loss=93.743, grad_norm=3.911, loss_scale=1.000, learning_rate=1.749e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 02:26:35,144 (deepspeed_trainer:226) INFO: 6epoch:train:3501-3600batch: iter_time=1.083e-04, loss_ctc=115.862, loss_att=92.751, acc=0.566, loss=99.708, grad_norm=4.344, loss_scale=1.000, learning_rate=1.748e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 02:27:13,367 (deepspeed_trainer:226) INFO: 6epoch:train:3601-3700batch: iter_time=1.066e-04, loss_ctc=104.405, loss_att=92.475, acc=0.562, loss=96.023, grad_norm=3.609, loss_scale=1.000, learning_rate=1.747e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 02:27:36,577 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-29 02:28:03,319 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 02:28:09,179 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 02:28:09,179 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-29 02:28:09,181 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 02:28:28,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:36,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:29,950] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:37,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:37,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:37,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:38,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:38,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:38,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:28:38,932] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:15,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:16,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:16,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:17,543] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:17,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:25,023] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:18,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:18,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:26,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:26,975] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:27,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:27,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:28,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:28,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:20,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:29:28,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:02,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:04,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:04,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:05,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:05,575] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:05,870] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:06,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:13,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:16,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:16,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:17,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:17,965] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:18,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:18,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:11,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:18,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:50,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:51,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:52,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:52,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:53,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:53,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:30:54,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:02,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:05,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:06,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:06,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:06,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:07,224] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:07,784] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:08,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:31:01,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 02:31:54,349 (deepspeed_trainer:226) INFO: 6epoch:train:3701-3800batch: iter_time=2.415, loss_ctc=88.166, loss_att=71.019, acc=0.582, loss=76.150, grad_norm=3.036, loss_scale=1.000, learning_rate=1.746e-04, step_time=0.394 [cnode7-008:0/16] 2024-11-29 02:32:33,590 (deepspeed_trainer:226) INFO: 6epoch:train:3801-3900batch: iter_time=1.143e-04, loss_ctc=116.562, loss_att=102.915, acc=0.547, loss=106.986, grad_norm=3.951, loss_scale=1.000, learning_rate=1.745e-04, step_time=0.392 [2024-11-29 02:33:11,614] [INFO] [logging.py:129:log_dist] [Rank 0] step=79000, skipped=0, lr=[np.float64(0.00017429676499164677)], mom=[[0.9, 0.98]] [2024-11-29 02:33:11,615] [INFO] [timer.py:264:stop] epoch=0/micro_step=79000/global_step=79000, RunningAvgSamplesPerSec=42.95234646628408, CurrSamplesPerSec=45.50959281904836, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:33:11,616 (deepspeed_trainer:226) INFO: 6epoch:train:3901-4000batch: iter_time=1.113e-04, loss_ctc=100.884, loss_att=85.541, acc=0.578, loss=90.123, grad_norm=3.508, loss_scale=1.000, learning_rate=1.744e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 02:33:50,330 (deepspeed_trainer:226) INFO: 6epoch:train:4001-4100batch: iter_time=1.123e-04, loss_ctc=93.226, loss_att=80.735, acc=0.578, loss=84.439, grad_norm=3.293, loss_scale=1.000, learning_rate=1.742e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 02:34:28,186 (deepspeed_trainer:226) INFO: 6epoch:train:4101-4200batch: iter_time=1.213e-04, loss_ctc=99.749, loss_att=87.479, acc=0.573, loss=91.132, grad_norm=3.243, loss_scale=1.000, learning_rate=1.741e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 02:35:05,867 (deepspeed_trainer:226) INFO: 6epoch:train:4201-4300batch: iter_time=1.189e-04, loss_ctc=119.733, loss_att=107.718, acc=0.567, loss=111.314, grad_norm=4.119, loss_scale=1.000, learning_rate=1.740e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 02:35:43,085 (deepspeed_trainer:226) INFO: 6epoch:train:4301-4400batch: iter_time=1.067e-04, loss_ctc=108.351, loss_att=87.355, acc=0.575, loss=93.669, grad_norm=4.005, loss_scale=1.000, learning_rate=1.739e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:36:20,275 (deepspeed_trainer:226) INFO: 6epoch:train:4401-4500batch: iter_time=1.093e-04, loss_ctc=110.173, loss_att=102.529, acc=0.557, loss=104.794, grad_norm=3.919, loss_scale=1.000, learning_rate=1.738e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:36:58,353 (deepspeed_trainer:226) INFO: 6epoch:train:4501-4600batch: iter_time=1.083e-04, loss_ctc=108.855, loss_att=99.147, acc=0.566, loss=102.026, grad_norm=3.828, loss_scale=1.000, learning_rate=1.737e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 02:37:35,785 (deepspeed_trainer:226) INFO: 6epoch:train:4601-4700batch: iter_time=1.080e-04, loss_ctc=89.024, loss_att=77.711, acc=0.582, loss=81.136, grad_norm=3.131, loss_scale=1.000, learning_rate=1.736e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:38:13,358 (deepspeed_trainer:226) INFO: 6epoch:train:4701-4800batch: iter_time=1.082e-04, loss_ctc=98.931, loss_att=79.858, acc=0.584, loss=85.588, grad_norm=3.687, loss_scale=1.000, learning_rate=1.735e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 02:38:50,595 (deepspeed_trainer:226) INFO: 6epoch:train:4801-4900batch: iter_time=1.075e-04, loss_ctc=82.595, loss_att=68.771, acc=0.591, loss=72.909, grad_norm=3.252, loss_scale=1.000, learning_rate=1.734e-04, step_time=0.372 [2024-11-29 02:39:28,013] [INFO] [logging.py:129:log_dist] [Rank 0] step=80000, skipped=0, lr=[np.float64(0.00017320399823528165)], mom=[[0.9, 0.98]] [2024-11-29 02:39:28,014] [INFO] [timer.py:264:stop] epoch=0/micro_step=80000/global_step=80000, RunningAvgSamplesPerSec=42.95737132578197, CurrSamplesPerSec=42.89182885283403, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:39:28,015 (deepspeed_trainer:226) INFO: 6epoch:train:4901-5000batch: iter_time=1.074e-04, loss_ctc=104.691, loss_att=84.613, acc=0.590, loss=90.603, grad_norm=3.684, loss_scale=1.000, learning_rate=1.733e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 02:40:06,431 (deepspeed_trainer:226) INFO: 6epoch:train:5001-5100batch: iter_time=1.091e-04, loss_ctc=98.584, loss_att=84.045, acc=0.587, loss=88.399, grad_norm=3.428, loss_scale=1.000, learning_rate=1.731e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 02:40:45,044 (deepspeed_trainer:226) INFO: 6epoch:train:5101-5200batch: iter_time=1.090e-04, loss_ctc=95.186, loss_att=82.037, acc=0.595, loss=85.979, grad_norm=3.150, loss_scale=1.000, learning_rate=1.730e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 02:41:23,601 (deepspeed_trainer:226) INFO: 6epoch:train:5201-5300batch: iter_time=1.086e-04, loss_ctc=115.572, loss_att=102.092, acc=0.568, loss=106.125, grad_norm=3.934, loss_scale=1.000, learning_rate=1.729e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 02:42:02,412 (deepspeed_trainer:226) INFO: 6epoch:train:5301-5400batch: iter_time=1.076e-04, loss_ctc=95.464, loss_att=82.374, acc=0.579, loss=86.291, grad_norm=3.180, loss_scale=1.000, learning_rate=1.728e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 02:42:40,686 (deepspeed_trainer:226) INFO: 6epoch:train:5401-5500batch: iter_time=1.097e-04, loss_ctc=113.686, loss_att=94.994, acc=0.571, loss=100.595, grad_norm=4.452, loss_scale=1.000, learning_rate=1.727e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 02:43:17,935 (deepspeed_trainer:226) INFO: 6epoch:train:5501-5600batch: iter_time=1.123e-04, loss_ctc=101.060, loss_att=89.959, acc=0.573, loss=93.320, grad_norm=3.453, loss_scale=1.000, learning_rate=1.726e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:43:30,833 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-29 02:43:58,128 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 02:44:04,029 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 02:44:04,029 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-29 02:44:04,032 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 02:44:22,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:30,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:22,859] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:23,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:31,070] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:23,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:31,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:24,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:31,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:32,213] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:32,477] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:24,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:24,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:25,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:32,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:44:33,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:11,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:19,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:12,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:12,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:12,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:20,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:20,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:13,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:13,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:13,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:13,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:21,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:22,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:22,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:22,804] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:45:22,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:00,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:00,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:08,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:00,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:01,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:01,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:02,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:02,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:02,768] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:10,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:10,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:10,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:11,503] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:12,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:14,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:16,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:48,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:56,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:49,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:49,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:50,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:51,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:51,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:59,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:59,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:59,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:47:00,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:53,064] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:46:54,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:47:02,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:47:04,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 02:47:07,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 02:48:01,446 (deepspeed_trainer:226) INFO: 6epoch:train:5601-5700batch: iter_time=2.439, loss_ctc=94.450, loss_att=74.237, acc=0.591, loss=80.304, grad_norm=3.208, loss_scale=1.000, learning_rate=1.725e-04, step_time=0.396 [cnode7-008:0/16] 2024-11-29 02:48:39,360 (deepspeed_trainer:226) INFO: 6epoch:train:5701-5800batch: iter_time=1.076e-04, loss_ctc=111.509, loss_att=95.276, acc=0.552, loss=100.103, grad_norm=3.752, loss_scale=1.000, learning_rate=1.724e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:49:16,240 (deepspeed_trainer:226) INFO: 6epoch:train:5801-5900batch: iter_time=1.129e-04, loss_ctc=98.574, loss_att=83.951, acc=0.577, loss=88.337, grad_norm=3.417, loss_scale=1.000, learning_rate=1.723e-04, step_time=0.368 [2024-11-29 02:49:53,538] [INFO] [logging.py:129:log_dist] [Rank 0] step=81000, skipped=0, lr=[np.float64(0.000172131530627741)], mom=[[0.9, 0.98]] [2024-11-29 02:49:53,539] [INFO] [timer.py:264:stop] epoch=0/micro_step=81000/global_step=81000, RunningAvgSamplesPerSec=42.955097652889194, CurrSamplesPerSec=42.61602171259288, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:49:53,540 (deepspeed_trainer:226) INFO: 6epoch:train:5901-6000batch: iter_time=1.099e-04, loss_ctc=95.820, loss_att=82.511, acc=0.575, loss=86.487, grad_norm=3.157, loss_scale=1.000, learning_rate=1.722e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:50:31,562 (deepspeed_trainer:226) INFO: 6epoch:train:6001-6100batch: iter_time=1.102e-04, loss_ctc=99.112, loss_att=89.328, acc=0.583, loss=92.274, grad_norm=3.273, loss_scale=1.000, learning_rate=1.721e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 02:51:09,717 (deepspeed_trainer:226) INFO: 6epoch:train:6101-6200batch: iter_time=1.091e-04, loss_ctc=111.419, loss_att=97.776, acc=0.570, loss=101.826, grad_norm=3.508, loss_scale=1.000, learning_rate=1.720e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 02:51:46,890 (deepspeed_trainer:226) INFO: 6epoch:train:6201-6300batch: iter_time=1.119e-04, loss_ctc=111.414, loss_att=97.779, acc=0.561, loss=101.889, grad_norm=4.016, loss_scale=1.000, learning_rate=1.719e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:52:24,933 (deepspeed_trainer:226) INFO: 6epoch:train:6301-6400batch: iter_time=1.106e-04, loss_ctc=109.687, loss_att=98.726, acc=0.556, loss=101.993, grad_norm=3.753, loss_scale=1.000, learning_rate=1.718e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 02:53:02,029 (deepspeed_trainer:226) INFO: 6epoch:train:6401-6500batch: iter_time=1.130e-04, loss_ctc=96.039, loss_att=84.777, acc=0.579, loss=88.141, grad_norm=3.350, loss_scale=1.000, learning_rate=1.717e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:53:38,634 (deepspeed_trainer:226) INFO: 6epoch:train:6501-6600batch: iter_time=1.109e-04, loss_ctc=86.396, loss_att=75.326, acc=0.585, loss=78.654, grad_norm=3.248, loss_scale=1.000, learning_rate=1.715e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 02:54:15,291 (deepspeed_trainer:226) INFO: 6epoch:train:6601-6700batch: iter_time=1.114e-04, loss_ctc=96.774, loss_att=75.265, acc=0.589, loss=81.725, grad_norm=3.595, loss_scale=1.000, learning_rate=1.714e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 02:54:51,940 (deepspeed_trainer:226) INFO: 6epoch:train:6701-6800batch: iter_time=1.098e-04, loss_ctc=91.732, loss_att=72.717, acc=0.592, loss=78.435, grad_norm=3.540, loss_scale=1.000, learning_rate=1.713e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 02:55:28,791 (deepspeed_trainer:226) INFO: 6epoch:train:6801-6900batch: iter_time=1.097e-04, loss_ctc=100.458, loss_att=80.798, acc=0.599, loss=86.693, grad_norm=3.569, loss_scale=1.000, learning_rate=1.712e-04, step_time=0.368 [2024-11-29 02:56:06,022] [INFO] [logging.py:129:log_dist] [Rank 0] step=82000, skipped=0, lr=[np.float64(0.0001710787413937834)], mom=[[0.9, 0.98]] [2024-11-29 02:56:06,023] [INFO] [timer.py:264:stop] epoch=0/micro_step=82000/global_step=82000, RunningAvgSamplesPerSec=42.96563447470605, CurrSamplesPerSec=43.53596111005854, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 02:56:06,024 (deepspeed_trainer:226) INFO: 6epoch:train:6901-7000batch: iter_time=1.117e-04, loss_ctc=93.433, loss_att=75.993, acc=0.587, loss=81.241, grad_norm=3.647, loss_scale=1.000, learning_rate=1.711e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 02:56:44,216 (deepspeed_trainer:226) INFO: 6epoch:train:7001-7100batch: iter_time=1.109e-04, loss_ctc=102.053, loss_att=86.197, acc=0.585, loss=90.945, grad_norm=3.896, loss_scale=1.000, learning_rate=1.710e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 02:57:22,295 (deepspeed_trainer:226) INFO: 6epoch:train:7101-7200batch: iter_time=1.118e-04, loss_ctc=108.830, loss_att=99.709, acc=0.562, loss=102.456, grad_norm=3.872, loss_scale=1.000, learning_rate=1.709e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 02:57:59,454 (deepspeed_trainer:226) INFO: 6epoch:train:7201-7300batch: iter_time=1.091e-04, loss_ctc=94.006, loss_att=74.472, acc=0.590, loss=80.337, grad_norm=3.631, loss_scale=1.000, learning_rate=1.708e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 02:58:38,368 (deepspeed_trainer:226) INFO: 6epoch:train:7301-7400batch: iter_time=1.134e-04, loss_ctc=109.653, loss_att=96.357, acc=0.567, loss=100.330, grad_norm=3.956, loss_scale=1.000, learning_rate=1.707e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-29 02:59:16,354 (deepspeed_trainer:226) INFO: 6epoch:train:7401-7500batch: iter_time=1.064e-04, loss_ctc=100.241, loss_att=80.448, acc=0.583, loss=86.366, grad_norm=3.486, loss_scale=1.000, learning_rate=1.706e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 02:59:20,453 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-29 02:59:47,161 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 02:59:52,911 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 02:59:52,911 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-29 02:59:52,914 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 03:00:21,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:21,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:21,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:21,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:13,768] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:13,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:14,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:22,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:22,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:22,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:00:22,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:09,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:09,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:09,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:10,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:02,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:02,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:10,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:02,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:03,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:03,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:03,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:03,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:11,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:03,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:12,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:13,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:59,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:51,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:51,361] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:52,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:00,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:52,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:00,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:00,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:52,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:52,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:52,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:01:53,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:01,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:01,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:04,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:05,224] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:38,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:47,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:40,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:40,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:40,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:40,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:49,159] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:49,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:41,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:49,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:41,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:50,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:43,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:52,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:52,899] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:02:56,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 03:04:01,440 (deepspeed_trainer:226) INFO: 6epoch:train:7501-7600batch: iter_time=2.437, loss_ctc=93.786, loss_att=78.716, acc=0.582, loss=83.232, grad_norm=3.853, loss_scale=1.000, learning_rate=1.705e-04, step_time=0.414 [cnode7-008:0/16] 2024-11-29 03:04:39,148 (deepspeed_trainer:226) INFO: 6epoch:train:7601-7700batch: iter_time=1.215e-04, loss_ctc=107.600, loss_att=85.046, acc=0.573, loss=91.823, grad_norm=3.928, loss_scale=1.000, learning_rate=1.704e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:05:16,469 (deepspeed_trainer:226) INFO: 6epoch:train:7701-7800batch: iter_time=1.220e-04, loss_ctc=93.586, loss_att=81.963, acc=0.585, loss=85.451, grad_norm=3.497, loss_scale=1.000, learning_rate=1.703e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 03:05:53,756 (deepspeed_trainer:226) INFO: 6epoch:train:7801-7900batch: iter_time=1.301e-04, loss_ctc=96.377, loss_att=76.656, acc=0.591, loss=82.587, grad_norm=3.456, loss_scale=1.000, learning_rate=1.702e-04, step_time=0.373 [2024-11-29 03:06:31,988] [INFO] [logging.py:129:log_dist] [Rank 0] step=83000, skipped=0, lr=[np.float64(0.0001700450360149393)], mom=[[0.9, 0.98]] [2024-11-29 03:06:31,989] [INFO] [timer.py:264:stop] epoch=0/micro_step=83000/global_step=83000, RunningAvgSamplesPerSec=42.962303427573076, CurrSamplesPerSec=43.4331444752799, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:06:31,990 (deepspeed_trainer:226) INFO: 6epoch:train:7901-8000batch: iter_time=1.318e-04, loss_ctc=104.395, loss_att=93.016, acc=0.581, loss=96.419, grad_norm=3.735, loss_scale=1.000, learning_rate=1.701e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 03:07:11,133 (deepspeed_trainer:226) INFO: 6epoch:train:8001-8100batch: iter_time=1.409e-04, loss_ctc=116.086, loss_att=97.723, acc=0.569, loss=103.232, grad_norm=4.251, loss_scale=1.000, learning_rate=1.700e-04, step_time=0.391 [cnode7-008:0/16] 2024-11-29 03:07:48,657 (deepspeed_trainer:226) INFO: 6epoch:train:8101-8200batch: iter_time=1.278e-04, loss_ctc=101.190, loss_att=92.651, acc=0.566, loss=95.176, grad_norm=3.719, loss_scale=1.000, learning_rate=1.699e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:08:26,247 (deepspeed_trainer:226) INFO: 6epoch:train:8201-8300batch: iter_time=1.227e-04, loss_ctc=106.697, loss_att=93.926, acc=0.568, loss=97.711, grad_norm=3.639, loss_scale=1.000, learning_rate=1.698e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:09:05,073 (deepspeed_trainer:226) INFO: 6epoch:train:8301-8400batch: iter_time=1.099e-04, loss_ctc=97.505, loss_att=86.207, acc=0.587, loss=89.593, grad_norm=3.253, loss_scale=1.000, learning_rate=1.697e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 03:09:43,455 (deepspeed_trainer:226) INFO: 6epoch:train:8401-8500batch: iter_time=1.116e-04, loss_ctc=81.321, loss_att=70.336, acc=0.591, loss=73.573, grad_norm=3.121, loss_scale=1.000, learning_rate=1.696e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 03:10:21,360 (deepspeed_trainer:226) INFO: 6epoch:train:8501-8600batch: iter_time=1.128e-04, loss_ctc=98.700, loss_att=76.682, acc=0.599, loss=83.261, grad_norm=3.348, loss_scale=1.000, learning_rate=1.695e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 03:10:59,072 (deepspeed_trainer:226) INFO: 6epoch:train:8601-8700batch: iter_time=1.114e-04, loss_ctc=91.252, loss_att=71.972, acc=0.592, loss=77.783, grad_norm=3.384, loss_scale=1.000, learning_rate=1.694e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:11:37,549 (deepspeed_trainer:226) INFO: 6epoch:train:8701-8800batch: iter_time=1.091e-04, loss_ctc=96.243, loss_att=74.896, acc=0.603, loss=81.331, grad_norm=3.534, loss_scale=1.000, learning_rate=1.693e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 03:12:15,142 (deepspeed_trainer:226) INFO: 6epoch:train:8801-8900batch: iter_time=1.112e-04, loss_ctc=92.791, loss_att=81.733, acc=0.574, loss=85.032, grad_norm=3.544, loss_scale=1.000, learning_rate=1.692e-04, step_time=0.375 [2024-11-29 03:12:52,628] [INFO] [logging.py:129:log_dist] [Rank 0] step=84000, skipped=0, lr=[np.float64(0.00016902984481866905)], mom=[[0.9, 0.98]] [2024-11-29 03:12:52,628] [INFO] [timer.py:264:stop] epoch=0/micro_step=84000/global_step=84000, RunningAvgSamplesPerSec=42.961097821610345, CurrSamplesPerSec=46.133038814809595, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:12:52,630 (deepspeed_trainer:226) INFO: 6epoch:train:8901-9000batch: iter_time=1.102e-04, loss_ctc=108.185, loss_att=86.966, acc=0.595, loss=93.263, grad_norm=3.972, loss_scale=1.000, learning_rate=1.691e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:13:29,936 (deepspeed_trainer:226) INFO: 6epoch:train:9001-9100batch: iter_time=1.116e-04, loss_ctc=100.976, loss_att=89.493, acc=0.574, loss=92.969, grad_norm=3.639, loss_scale=1.000, learning_rate=1.690e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 03:14:07,150 (deepspeed_trainer:226) INFO: 6epoch:train:9101-9200batch: iter_time=1.087e-04, loss_ctc=103.664, loss_att=81.416, acc=0.587, loss=88.070, grad_norm=3.914, loss_scale=1.000, learning_rate=1.689e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 03:14:44,486 (deepspeed_trainer:226) INFO: 6epoch:train:9201-9300batch: iter_time=1.108e-04, loss_ctc=105.134, loss_att=91.536, acc=0.571, loss=95.567, grad_norm=4.332, loss_scale=1.000, learning_rate=1.688e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 03:15:16,827 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-29 03:15:44,509 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 03:15:50,004 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 03:15:50,004 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-29 03:15:50,007 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 03:16:08,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:08,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:08,920] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:17,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:09,450] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:09,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:10,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:17,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:10,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:10,424] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:18,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:19,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:19,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:19,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:19,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:19,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:57,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:06,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:58,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:58,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:58,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:59,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:06,943] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:59,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:16:59,501] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:07,589] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:08,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:08,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:08,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:09,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:02,243] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:11,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:45,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:46,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:46,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:54,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:46,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:47,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:47,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:48,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:56,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:56,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:57,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:58,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:58,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:58,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:17:52,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:01,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:33,182] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:34,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:35,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:43,495] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:36,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:36,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:36,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:44,950] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:45,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:46,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:47,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:39,714] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:47,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:48,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:49,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:18:43,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 03:19:26,341 (deepspeed_trainer:226) INFO: 6epoch:train:9301-9400batch: iter_time=2.413, loss_ctc=91.441, loss_att=71.950, acc=0.597, loss=77.796, grad_norm=3.070, loss_scale=1.000, learning_rate=1.687e-04, step_time=0.405 [cnode7-008:0/16] 2024-11-29 03:20:05,415 (deepspeed_trainer:226) INFO: 6epoch:train:9401-9500batch: iter_time=1.098e-04, loss_ctc=102.000, loss_att=91.207, acc=0.575, loss=94.434, grad_norm=3.786, loss_scale=1.000, learning_rate=1.686e-04, step_time=0.390 [cnode7-008:0/16] 2024-11-29 03:20:43,499 (deepspeed_trainer:226) INFO: 6epoch:train:9501-9600batch: iter_time=1.101e-04, loss_ctc=102.977, loss_att=85.662, acc=0.579, loss=90.887, grad_norm=3.682, loss_scale=1.000, learning_rate=1.685e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 03:21:21,125 (deepspeed_trainer:226) INFO: 6epoch:train:9601-9700batch: iter_time=1.171e-04, loss_ctc=91.731, loss_att=81.018, acc=0.596, loss=84.256, grad_norm=3.385, loss_scale=1.000, learning_rate=1.684e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:21:59,404 (deepspeed_trainer:226) INFO: 6epoch:train:9701-9800batch: iter_time=1.113e-04, loss_ctc=94.963, loss_att=82.419, acc=0.592, loss=86.181, grad_norm=3.064, loss_scale=1.000, learning_rate=1.683e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 03:22:36,944 (deepspeed_trainer:226) INFO: 6epoch:train:9801-9900batch: iter_time=1.114e-04, loss_ctc=108.269, loss_att=92.075, acc=0.586, loss=96.933, grad_norm=3.557, loss_scale=1.000, learning_rate=1.682e-04, step_time=0.375 [2024-11-29 03:23:14,604] [INFO] [logging.py:129:log_dist] [Rank 0] step=85000, skipped=0, lr=[np.float64(0.00016803262165908207)], mom=[[0.9, 0.98]] [2024-11-29 03:23:14,605] [INFO] [timer.py:264:stop] epoch=0/micro_step=85000/global_step=85000, RunningAvgSamplesPerSec=42.960006348356664, CurrSamplesPerSec=45.69586443745039, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:23:14,606 (deepspeed_trainer:226) INFO: 6epoch:train:9901-10000batch: iter_time=1.096e-04, loss_ctc=109.754, loss_att=95.368, acc=0.579, loss=99.637, grad_norm=3.764, loss_scale=1.000, learning_rate=1.681e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:23:51,728 (deepspeed_trainer:226) INFO: 6epoch:train:10001-10100batch: iter_time=1.086e-04, loss_ctc=102.940, loss_att=94.212, acc=0.577, loss=96.854, grad_norm=3.718, loss_scale=1.000, learning_rate=1.680e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 03:24:28,637 (deepspeed_trainer:226) INFO: 6epoch:train:10101-10200batch: iter_time=1.106e-04, loss_ctc=106.199, loss_att=96.286, acc=0.580, loss=99.211, grad_norm=3.701, loss_scale=1.000, learning_rate=1.679e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 03:25:05,222 (deepspeed_trainer:226) INFO: 6epoch:train:10201-10300batch: iter_time=1.093e-04, loss_ctc=91.425, loss_att=82.319, acc=0.591, loss=85.043, grad_norm=3.652, loss_scale=1.000, learning_rate=1.678e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 03:25:42,087 (deepspeed_trainer:226) INFO: 6epoch:train:10301-10400batch: iter_time=1.070e-04, loss_ctc=85.754, loss_att=69.509, acc=0.602, loss=74.396, grad_norm=3.655, loss_scale=1.000, learning_rate=1.677e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 03:26:19,284 (deepspeed_trainer:226) INFO: 6epoch:train:10401-10500batch: iter_time=1.160e-04, loss_ctc=93.414, loss_att=76.758, acc=0.603, loss=81.771, grad_norm=3.802, loss_scale=1.000, learning_rate=1.676e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 03:26:56,275 (deepspeed_trainer:226) INFO: 6epoch:train:10501-10600batch: iter_time=1.118e-04, loss_ctc=92.674, loss_att=71.862, acc=0.601, loss=78.136, grad_norm=3.917, loss_scale=1.000, learning_rate=1.675e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 03:27:33,314 (deepspeed_trainer:226) INFO: 6epoch:train:10601-10700batch: iter_time=1.119e-04, loss_ctc=95.670, loss_att=81.256, acc=0.603, loss=85.551, grad_norm=3.649, loss_scale=1.000, learning_rate=1.674e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 03:28:10,792 (deepspeed_trainer:226) INFO: 6epoch:train:10701-10800batch: iter_time=1.093e-04, loss_ctc=87.799, loss_att=77.734, acc=0.604, loss=80.750, grad_norm=3.311, loss_scale=1.000, learning_rate=1.673e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:28:50,427 (deepspeed_trainer:226) INFO: 6epoch:train:10801-10900batch: iter_time=1.094e-04, loss_ctc=112.682, loss_att=92.142, acc=0.592, loss=98.307, grad_norm=4.358, loss_scale=1.000, learning_rate=1.672e-04, step_time=0.396 [2024-11-29 03:29:28,120] [INFO] [logging.py:129:log_dist] [Rank 0] step=86000, skipped=0, lr=[np.float64(0.00016705284268227726)], mom=[[0.9, 0.98]] [2024-11-29 03:29:28,120] [INFO] [timer.py:264:stop] epoch=0/micro_step=86000/global_step=86000, RunningAvgSamplesPerSec=42.968493365948135, CurrSamplesPerSec=46.31496438553616, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:29:28,121 (deepspeed_trainer:226) INFO: 6epoch:train:10901-11000batch: iter_time=1.161e-04, loss_ctc=95.611, loss_att=86.716, acc=0.588, loss=89.412, grad_norm=3.627, loss_scale=1.000, learning_rate=1.671e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:30:05,762 (deepspeed_trainer:226) INFO: 6epoch:train:11001-11100batch: iter_time=1.119e-04, loss_ctc=109.218, loss_att=88.471, acc=0.590, loss=94.681, grad_norm=4.721, loss_scale=1.000, learning_rate=1.670e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:30:44,094 (deepspeed_trainer:226) INFO: 6epoch:train:11101-11200batch: iter_time=1.111e-04, loss_ctc=98.595, loss_att=87.781, acc=0.586, loss=91.022, grad_norm=3.885, loss_scale=1.000, learning_rate=1.669e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 03:31:07,255 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-29 03:31:34,449 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 03:31:40,325 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 03:31:40,325 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-29 03:31:40,327 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 03:31:58,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:31:58,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:07,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:07,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:31:59,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:07,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:00,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:00,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:00,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:00,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:00,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:08,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:08,945] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:09,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:09,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:09,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:47,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:47,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:47,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:56,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:56,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:48,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:56,552] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:49,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:49,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:49,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:57,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:49,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:58,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:58,488] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:59,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:32:59,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:34,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:35,409] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:36,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:44,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:44,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:45,023] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:37,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:37,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:45,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:38,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:38,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:46,254] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:38,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:47,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:47,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:33:47,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:22,613] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:23,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:32,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:24,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:32,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:24,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:32,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:25,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:26,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:34,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:26,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:34,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:26,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:34,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:34,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:34:37,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 03:35:23,559 (deepspeed_trainer:226) INFO: 6epoch:train:11201-11300batch: iter_time=2.389, loss_ctc=84.006, loss_att=67.153, acc=0.607, loss=72.238, grad_norm=3.256, loss_scale=1.000, learning_rate=1.668e-04, step_time=0.405 [cnode7-008:0/16] 2024-11-29 03:36:01,783 (deepspeed_trainer:226) INFO: 6epoch:train:11301-11400batch: iter_time=1.100e-04, loss_ctc=110.787, loss_att=95.243, acc=0.571, loss=99.918, grad_norm=3.922, loss_scale=1.000, learning_rate=1.667e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 03:36:39,601 (deepspeed_trainer:226) INFO: 6epoch:train:11401-11500batch: iter_time=1.044e-04, loss_ctc=95.079, loss_att=78.689, acc=0.603, loss=83.609, grad_norm=3.269, loss_scale=1.000, learning_rate=1.666e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 03:37:16,796 (deepspeed_trainer:226) INFO: 6epoch:train:11501-11600batch: iter_time=1.099e-04, loss_ctc=88.265, loss_att=75.354, acc=0.597, loss=79.231, grad_norm=3.270, loss_scale=1.000, learning_rate=1.665e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 03:37:53,816 (deepspeed_trainer:226) INFO: 6epoch:train:11601-11700batch: iter_time=1.084e-04, loss_ctc=95.001, loss_att=81.313, acc=0.595, loss=85.432, grad_norm=3.453, loss_scale=1.000, learning_rate=1.664e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 03:38:31,407 (deepspeed_trainer:226) INFO: 6epoch:train:11701-11800batch: iter_time=1.094e-04, loss_ctc=112.469, loss_att=100.506, acc=0.587, loss=104.129, grad_norm=3.949, loss_scale=1.000, learning_rate=1.663e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:39:11,182 (deepspeed_trainer:226) INFO: 6epoch:train:11801-11900batch: iter_time=1.060e-04, loss_ctc=104.252, loss_att=82.539, acc=0.593, loss=89.067, grad_norm=3.979, loss_scale=1.000, learning_rate=1.662e-04, step_time=0.397 [2024-11-29 03:39:48,804] [INFO] [logging.py:129:log_dist] [Rank 0] step=87000, skipped=0, lr=[np.float64(0.00016609000516996326)], mom=[[0.9, 0.98]] [2024-11-29 03:39:48,805] [INFO] [timer.py:264:stop] epoch=0/micro_step=87000/global_step=87000, RunningAvgSamplesPerSec=42.96568575917651, CurrSamplesPerSec=44.195572472902334, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:39:48,806 (deepspeed_trainer:226) INFO: 6epoch:train:11901-12000batch: iter_time=1.085e-04, loss_ctc=105.147, loss_att=96.887, acc=0.576, loss=99.387, grad_norm=4.007, loss_scale=1.000, learning_rate=1.661e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:40:26,406 (deepspeed_trainer:226) INFO: 6epoch:train:12001-12100batch: iter_time=1.100e-04, loss_ctc=103.626, loss_att=93.221, acc=0.586, loss=96.356, grad_norm=3.825, loss_scale=1.000, learning_rate=1.660e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:41:03,943 (deepspeed_trainer:226) INFO: 6epoch:train:12101-12200batch: iter_time=1.083e-04, loss_ctc=85.154, loss_att=74.272, acc=0.598, loss=77.537, grad_norm=3.079, loss_scale=1.000, learning_rate=1.659e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:41:41,661 (deepspeed_trainer:226) INFO: 6epoch:train:12201-12300batch: iter_time=1.110e-04, loss_ctc=95.286, loss_att=75.476, acc=0.604, loss=81.411, grad_norm=3.631, loss_scale=1.000, learning_rate=1.659e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:42:18,236 (deepspeed_trainer:226) INFO: 6epoch:train:12301-12400batch: iter_time=1.084e-04, loss_ctc=79.194, loss_att=64.590, acc=0.608, loss=68.953, grad_norm=3.473, loss_scale=1.000, learning_rate=1.658e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 03:42:55,280 (deepspeed_trainer:226) INFO: 6epoch:train:12401-12500batch: iter_time=1.087e-04, loss_ctc=98.441, loss_att=78.889, acc=0.609, loss=84.731, grad_norm=3.631, loss_scale=1.000, learning_rate=1.657e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 03:43:32,179 (deepspeed_trainer:226) INFO: 6epoch:train:12501-12600batch: iter_time=1.112e-04, loss_ctc=94.839, loss_att=79.381, acc=0.605, loss=84.029, grad_norm=3.458, loss_scale=1.000, learning_rate=1.656e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 03:44:09,398 (deepspeed_trainer:226) INFO: 6epoch:train:12601-12700batch: iter_time=1.083e-04, loss_ctc=91.692, loss_att=77.812, acc=0.612, loss=81.941, grad_norm=3.405, loss_scale=1.000, learning_rate=1.655e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 03:44:47,081 (deepspeed_trainer:226) INFO: 6epoch:train:12701-12800batch: iter_time=1.145e-04, loss_ctc=110.766, loss_att=95.489, acc=0.588, loss=100.066, grad_norm=4.304, loss_scale=1.000, learning_rate=1.654e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:45:24,245 (deepspeed_trainer:226) INFO: 6epoch:train:12801-12900batch: iter_time=1.108e-04, loss_ctc=91.211, loss_att=77.079, acc=0.603, loss=81.310, grad_norm=3.299, loss_scale=1.000, learning_rate=1.653e-04, step_time=0.371 [2024-11-29 03:46:02,417] [INFO] [logging.py:129:log_dist] [Rank 0] step=88000, skipped=0, lr=[np.float64(0.00016514362645556036)], mom=[[0.9, 0.98]] [2024-11-29 03:46:02,418] [INFO] [timer.py:264:stop] epoch=0/micro_step=88000/global_step=88000, RunningAvgSamplesPerSec=42.973809317881276, CurrSamplesPerSec=43.814145130864496, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:46:02,419 (deepspeed_trainer:226) INFO: 6epoch:train:12901-13000batch: iter_time=1.105e-04, loss_ctc=106.780, loss_att=89.504, acc=0.587, loss=94.667, grad_norm=4.445, loss_scale=1.000, learning_rate=1.652e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 03:46:39,408 (deepspeed_trainer:226) INFO: 6epoch:train:13001-13100batch: iter_time=1.109e-04, loss_ctc=96.252, loss_att=83.641, acc=0.594, loss=87.417, grad_norm=3.581, loss_scale=1.000, learning_rate=1.651e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 03:46:52,417 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-29 03:47:19,772 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 03:47:25,420 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 03:47:25,420 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-29 03:47:25,422 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 03:47:45,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:45,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:46,321] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:54,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:47,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:54,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:54,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:47,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:47,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:47,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:47,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:55,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:55,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:56,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:56,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:47:56,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:34,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:34,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:34,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:35,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:35,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:43,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:43,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:36,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:36,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:43,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:36,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:43,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:44,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:45,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:46,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:48:47,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:22,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:23,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:23,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:24,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:31,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:24,655] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:32,143] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:25,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:25,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:25,475] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:32,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:33,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:33,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:33,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:34,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:49:37,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:10,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:11,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:12,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:12,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:12,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:13,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:13,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:20,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:21,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:14,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:21,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:22,280] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:22,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:23,446] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:24,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 03:50:26,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 03:51:24,065 (deepspeed_trainer:226) INFO: 6epoch:train:13101-13200batch: iter_time=2.437, loss_ctc=89.858, loss_att=70.727, acc=0.605, loss=76.474, grad_norm=3.309, loss_scale=1.000, learning_rate=1.650e-04, step_time=0.410 [cnode7-008:0/16] 2024-11-29 03:52:02,443 (deepspeed_trainer:226) INFO: 6epoch:train:13201-13300batch: iter_time=1.113e-04, loss_ctc=106.566, loss_att=90.254, acc=0.569, loss=95.179, grad_norm=3.666, loss_scale=1.000, learning_rate=1.649e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 03:52:40,349 (deepspeed_trainer:226) INFO: 6epoch:train:13301-13400batch: iter_time=1.120e-04, loss_ctc=94.004, loss_att=79.506, acc=0.594, loss=83.844, grad_norm=3.461, loss_scale=1.000, learning_rate=1.648e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 03:53:17,309 (deepspeed_trainer:226) INFO: 6epoch:train:13401-13500batch: iter_time=1.178e-04, loss_ctc=91.492, loss_att=78.341, acc=0.593, loss=82.291, grad_norm=3.146, loss_scale=1.000, learning_rate=1.647e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 03:53:54,522 (deepspeed_trainer:226) INFO: 6epoch:train:13501-13600batch: iter_time=1.171e-04, loss_ctc=94.371, loss_att=84.667, acc=0.600, loss=87.574, grad_norm=3.189, loss_scale=1.000, learning_rate=1.646e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 03:54:31,999 (deepspeed_trainer:226) INFO: 6epoch:train:13601-13700batch: iter_time=1.177e-04, loss_ctc=106.230, loss_att=91.928, acc=0.586, loss=96.226, grad_norm=3.754, loss_scale=1.000, learning_rate=1.645e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 03:55:09,622 (deepspeed_trainer:226) INFO: 6epoch:train:13701-13800batch: iter_time=1.231e-04, loss_ctc=106.623, loss_att=92.536, acc=0.580, loss=96.762, grad_norm=4.046, loss_scale=1.000, learning_rate=1.644e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 03:55:47,111 (deepspeed_trainer:226) INFO: 6epoch:train:13801-13900batch: iter_time=1.139e-04, loss_ctc=104.438, loss_att=94.021, acc=0.573, loss=97.109, grad_norm=3.886, loss_scale=1.000, learning_rate=1.644e-04, step_time=0.375 [2024-11-29 03:56:24,820] [INFO] [logging.py:129:log_dist] [Rank 0] step=89000, skipped=0, lr=[np.float64(0.00016421324290747494)], mom=[[0.9, 0.98]] [2024-11-29 03:56:24,821] [INFO] [timer.py:264:stop] epoch=0/micro_step=89000/global_step=89000, RunningAvgSamplesPerSec=42.97551782169007, CurrSamplesPerSec=44.872898260295585, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 03:56:24,822 (deepspeed_trainer:226) INFO: 6epoch:train:13901-14000batch: iter_time=1.167e-04, loss_ctc=91.998, loss_att=79.934, acc=0.596, loss=83.553, grad_norm=3.446, loss_scale=1.000, learning_rate=1.643e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 03:57:02,351 (deepspeed_trainer:226) INFO: 6epoch:train:14001-14100batch: iter_time=1.086e-04, loss_ctc=81.859, loss_att=71.593, acc=0.602, loss=74.641, grad_norm=3.048, loss_scale=1.000, learning_rate=1.642e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 03:57:39,314 (deepspeed_trainer:226) INFO: 6epoch:train:14101-14200batch: iter_time=1.109e-04, loss_ctc=92.461, loss_att=71.244, acc=0.606, loss=77.636, grad_norm=3.345, loss_scale=1.000, learning_rate=1.641e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 03:58:15,978 (deepspeed_trainer:226) INFO: 6epoch:train:14201-14300batch: iter_time=1.077e-04, loss_ctc=88.282, loss_att=69.259, acc=0.606, loss=74.992, grad_norm=3.558, loss_scale=1.000, learning_rate=1.640e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 03:58:53,371 (deepspeed_trainer:226) INFO: 6epoch:train:14301-14400batch: iter_time=1.128e-04, loss_ctc=95.601, loss_att=76.108, acc=0.616, loss=81.968, grad_norm=3.264, loss_scale=1.000, learning_rate=1.639e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 03:59:31,867 (deepspeed_trainer:226) INFO: 6epoch:train:14401-14500batch: iter_time=1.099e-04, loss_ctc=89.869, loss_att=72.389, acc=0.604, loss=77.640, grad_norm=3.383, loss_scale=1.000, learning_rate=1.638e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 04:00:10,280 (deepspeed_trainer:226) INFO: 6epoch:train:14501-14600batch: iter_time=1.082e-04, loss_ctc=97.659, loss_att=81.806, acc=0.602, loss=86.554, grad_norm=4.260, loss_scale=1.000, learning_rate=1.637e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 04:00:48,620 (deepspeed_trainer:226) INFO: 6epoch:train:14601-14700batch: iter_time=1.085e-04, loss_ctc=103.892, loss_att=94.183, acc=0.578, loss=97.081, grad_norm=3.869, loss_scale=1.000, learning_rate=1.636e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 04:01:26,170 (deepspeed_trainer:226) INFO: 6epoch:train:14701-14800batch: iter_time=1.087e-04, loss_ctc=89.422, loss_att=71.165, acc=0.606, loss=76.670, grad_norm=3.774, loss_scale=1.000, learning_rate=1.635e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 04:02:03,602 (deepspeed_trainer:226) INFO: 6epoch:train:14801-14900batch: iter_time=1.112e-04, loss_ctc=104.392, loss_att=91.020, acc=0.582, loss=95.043, grad_norm=4.108, loss_scale=1.000, learning_rate=1.634e-04, step_time=0.374 [2024-11-29 04:02:40,547] [INFO] [logging.py:129:log_dist] [Rank 0] step=90000, skipped=0, lr=[np.float64(0.00016329840897468204)], mom=[[0.9, 0.98]] [2024-11-29 04:02:40,547] [INFO] [timer.py:264:stop] epoch=0/micro_step=90000/global_step=90000, RunningAvgSamplesPerSec=42.98073871195154, CurrSamplesPerSec=40.31610392757981, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 04:02:40,548 (deepspeed_trainer:226) INFO: 6epoch:train:14901-15000batch: iter_time=1.088e-04, loss_ctc=96.828, loss_att=76.631, acc=0.599, loss=82.686, grad_norm=3.469, loss_scale=1.000, learning_rate=1.633e-04, step_time=0.369 [2024-11-29 04:02:47,532] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:55,955] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:48,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:56,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:56,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:56,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:49,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:56,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:56,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:57,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:02:57,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:02,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:11,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:04,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:11,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:04,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:04,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:12,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:12,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:05,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:06,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:13,591] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:06,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:06,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:13,693] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:13,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:13,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:17,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:26,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:19,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:20,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:20,354] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:28,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:28,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:28,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:28,916] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:21,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:29,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:22,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:29,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:22,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:22,237] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:29,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:32,165] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:42,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:34,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:35,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:35,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:43,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:36,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:44,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:44,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:44,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:37,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:45,253] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:37,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:37,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:45,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:45,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:03:58,343] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 6 is about to be saved! [2024-11-29 04:03:58,373] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/mp_rank_00_model_states.pt [2024-11-29 04:03:58,373] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/mp_rank_00_model_states.pt... [2024-11-29 04:04:00,209] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/mp_rank_00_model_states.pt. [2024-11-29 04:04:00,364] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,365] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,833] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,365] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,834] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,834] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,834] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,367] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,367] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,836] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,837] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,370] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,838] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,370] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-29 04:03:52,839] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,372] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-29 04:04:00,885] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-29 04:04:00,885] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-29 04:04:00,885] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,434] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,434] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,434] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,460] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,460] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,460] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,088] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,088] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,088] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,111] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,111] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,112] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,112] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,118] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,119] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,589] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,589] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,589] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,124] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,124] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,124] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,621] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,621] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,621] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,155] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,155] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,155] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,626] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,626] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,626] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,636] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,636] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,636] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,171] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,172] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,172] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,655] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,655] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,655] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:04:01,197] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-29 04:04:01,198] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-29 04:04:01,198] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [2024-11-29 04:03:53,679] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-29 04:03:53,679] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_6/6/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-29 04:03:53,679] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 6 is ready now! [cnode7-008:0/16] 2024-11-29 04:04:01,221 (deepspeed_trainer:158) INFO: 6epoch results: [train] iter_time=0.129, loss_ctc=100.576, loss_att=85.801, acc=0.581, loss=90.232, grad_norm=3.635, loss_scale=1.000, learning_rate=1.707e-04, step_time=0.378, time=2 hours, 6 minutes and 50.83 seconds, total_count=90006, gpu_max_cached_mem_GB=28.338, [valid] loss_ctc=10.062, cer_ctc=0.314, loss_att=12.250, acc=0.845, cer=0.157, wer=1.000, loss=11.625, time=1 minute and 8.83 seconds, total_count=6, gpu_max_cached_mem_GB=28.338 [cnode7-008:0/16] 2024-11-29 04:04:03,642 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-29 04:04:30,096 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 04:04:35,802 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 04:04:35,802 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-29 04:04:35,804 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 04:04:48,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:49,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:49,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:49,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:50,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:50,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:58,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:50,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:51,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:59,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:59,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:59,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:04:59,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:00,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:00,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:00,955] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:36,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:36,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:37,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:38,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:38,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:46,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:46,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:39,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:39,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:39,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:47,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:47,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:48,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:48,165] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:48,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:05:48,475] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:23,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:24,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:25,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:25,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:26,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:26,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:26,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:27,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:35,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:35,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:36,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:37,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:37,144] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:37,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:37,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:06:37,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:10,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:11,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:13,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:13,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:14,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:14,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:24,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:16,943] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:24,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:24,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:25,591] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:18,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:26,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:26,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:26,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:07:26,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 04:08:32,476 (deepspeed_trainer:226) INFO: 7epoch:train:1-100batch: iter_time=2.308, loss_ctc=104.138, loss_att=78.061, acc=0.624, loss=85.904, grad_norm=3.669, loss_scale=1.000, learning_rate=1.633e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 04:09:10,326 (deepspeed_trainer:226) INFO: 7epoch:train:101-200batch: iter_time=1.137e-04, loss_ctc=98.173, loss_att=83.848, acc=0.599, loss=88.111, grad_norm=3.489, loss_scale=1.000, learning_rate=1.632e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 04:09:47,544 (deepspeed_trainer:226) INFO: 7epoch:train:201-300batch: iter_time=1.135e-04, loss_ctc=102.213, loss_att=93.719, acc=0.567, loss=96.255, grad_norm=3.848, loss_scale=1.000, learning_rate=1.631e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 04:10:24,339 (deepspeed_trainer:226) INFO: 7epoch:train:301-400batch: iter_time=1.051e-04, loss_ctc=94.627, loss_att=82.909, acc=0.591, loss=86.426, grad_norm=3.465, loss_scale=1.000, learning_rate=1.630e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 04:11:03,219 (deepspeed_trainer:226) INFO: 7epoch:train:401-500batch: iter_time=1.047e-04, loss_ctc=100.115, loss_att=76.624, acc=0.600, loss=83.661, grad_norm=3.735, loss_scale=1.000, learning_rate=1.629e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 04:11:41,359 (deepspeed_trainer:226) INFO: 7epoch:train:501-600batch: iter_time=1.056e-04, loss_ctc=92.034, loss_att=73.633, acc=0.608, loss=79.176, grad_norm=3.281, loss_scale=1.000, learning_rate=1.628e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 04:12:19,130 (deepspeed_trainer:226) INFO: 7epoch:train:601-700batch: iter_time=1.071e-04, loss_ctc=104.064, loss_att=83.119, acc=0.603, loss=89.409, grad_norm=3.727, loss_scale=1.000, learning_rate=1.627e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 04:12:56,814 (deepspeed_trainer:226) INFO: 7epoch:train:701-800batch: iter_time=1.087e-04, loss_ctc=106.220, loss_att=85.865, acc=0.575, loss=91.998, grad_norm=3.758, loss_scale=1.000, learning_rate=1.626e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 04:13:34,903 (deepspeed_trainer:226) INFO: 7epoch:train:801-900batch: iter_time=1.104e-04, loss_ctc=95.289, loss_att=73.125, acc=0.611, loss=79.768, grad_norm=3.372, loss_scale=1.000, learning_rate=1.625e-04, step_time=0.381 [2024-11-29 04:14:12,563] [INFO] [logging.py:129:log_dist] [Rank 0] step=91000, skipped=0, lr=[np.float64(0.00016239869629015234)], mom=[[0.9, 0.98]] [2024-11-29 04:14:12,564] [INFO] [timer.py:264:stop] epoch=0/micro_step=91000/global_step=91000, RunningAvgSamplesPerSec=42.982354293951985, CurrSamplesPerSec=44.81326838110071, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 04:14:12,566 (deepspeed_trainer:226) INFO: 7epoch:train:901-1000batch: iter_time=1.094e-04, loss_ctc=110.177, loss_att=100.502, acc=0.573, loss=103.427, grad_norm=3.945, loss_scale=1.000, learning_rate=1.624e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 04:14:49,761 (deepspeed_trainer:226) INFO: 7epoch:train:1001-1100batch: iter_time=1.105e-04, loss_ctc=94.379, loss_att=84.487, acc=0.588, loss=87.434, grad_norm=3.316, loss_scale=1.000, learning_rate=1.624e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 04:15:26,859 (deepspeed_trainer:226) INFO: 7epoch:train:1101-1200batch: iter_time=1.102e-04, loss_ctc=115.624, loss_att=97.456, acc=0.586, loss=102.932, grad_norm=3.850, loss_scale=1.000, learning_rate=1.623e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 04:16:03,961 (deepspeed_trainer:226) INFO: 7epoch:train:1201-1300batch: iter_time=1.117e-04, loss_ctc=107.619, loss_att=85.201, acc=0.574, loss=91.915, grad_norm=4.392, loss_scale=1.000, learning_rate=1.622e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 04:16:41,585 (deepspeed_trainer:226) INFO: 7epoch:train:1301-1400batch: iter_time=1.108e-04, loss_ctc=99.452, loss_att=88.051, acc=0.599, loss=91.427, grad_norm=4.021, loss_scale=1.000, learning_rate=1.621e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 04:17:18,295 (deepspeed_trainer:226) INFO: 7epoch:train:1401-1500batch: iter_time=1.118e-04, loss_ctc=93.688, loss_att=78.491, acc=0.586, loss=83.059, grad_norm=3.564, loss_scale=1.000, learning_rate=1.620e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 04:17:55,071 (deepspeed_trainer:226) INFO: 7epoch:train:1501-1600batch: iter_time=1.080e-04, loss_ctc=100.521, loss_att=83.607, acc=0.576, loss=88.693, grad_norm=4.394, loss_scale=1.000, learning_rate=1.619e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 04:18:32,623 (deepspeed_trainer:226) INFO: 7epoch:train:1601-1700batch: iter_time=1.085e-04, loss_ctc=99.060, loss_att=86.029, acc=0.584, loss=89.927, grad_norm=3.389, loss_scale=1.000, learning_rate=1.618e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 04:19:10,735 (deepspeed_trainer:226) INFO: 7epoch:train:1701-1800batch: iter_time=1.072e-04, loss_ctc=107.200, loss_att=84.219, acc=0.610, loss=91.138, grad_norm=3.923, loss_scale=1.000, learning_rate=1.617e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 04:19:44,455 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-29 04:20:11,467 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 04:20:17,268 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 04:20:17,268 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-29 04:20:17,270 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 04:20:39,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,362] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,540] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:39,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:48,673] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:48,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,159] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:20:49,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:29,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:29,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:30,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:30,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:30,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:31,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:31,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:31,687] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:42,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:43,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:43,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:43,916] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:43,989] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:44,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:44,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:21:45,648] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:18,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:18,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:19,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:19,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:20,005] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:20,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:20,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:20,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:31,447] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:32,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:32,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:33,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:34,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:34,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:34,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:22:36,442] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:05,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:06,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:08,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:08,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:09,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:09,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:09,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:09,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:18,971] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:20,700] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:21,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:22,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:22,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:22,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:23,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:23:23,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 04:24:00,437 (deepspeed_trainer:226) INFO: 7epoch:train:1801-1900batch: iter_time=2.504, loss_ctc=92.790, loss_att=73.327, acc=0.620, loss=79.162, grad_norm=3.327, loss_scale=1.000, learning_rate=1.616e-04, step_time=0.393 [2024-11-29 04:24:37,891] [INFO] [logging.py:129:log_dist] [Rank 0] step=92000, skipped=0, lr=[np.float64(0.00016151369282802608)], mom=[[0.9, 0.98]] [2024-11-29 04:24:37,892] [INFO] [timer.py:264:stop] epoch=0/micro_step=92000/global_step=92000, RunningAvgSamplesPerSec=42.988040881341036, CurrSamplesPerSec=44.44389242975608, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 04:24:37,894 (deepspeed_trainer:226) INFO: 7epoch:train:1901-2000batch: iter_time=1.170e-04, loss_ctc=98.632, loss_att=77.239, acc=0.617, loss=83.665, grad_norm=3.559, loss_scale=1.000, learning_rate=1.616e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 04:25:15,297 (deepspeed_trainer:226) INFO: 7epoch:train:2001-2100batch: iter_time=1.075e-04, loss_ctc=106.296, loss_att=96.169, acc=0.599, loss=99.215, grad_norm=3.705, loss_scale=1.000, learning_rate=1.615e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 04:25:53,106 (deepspeed_trainer:226) INFO: 7epoch:train:2101-2200batch: iter_time=1.055e-04, loss_ctc=97.528, loss_att=89.501, acc=0.586, loss=91.999, grad_norm=3.697, loss_scale=1.000, learning_rate=1.614e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 04:26:29,994 (deepspeed_trainer:226) INFO: 7epoch:train:2201-2300batch: iter_time=1.061e-04, loss_ctc=97.347, loss_att=81.868, acc=0.587, loss=86.522, grad_norm=4.020, loss_scale=1.000, learning_rate=1.613e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 04:27:06,554 (deepspeed_trainer:226) INFO: 7epoch:train:2301-2400batch: iter_time=1.051e-04, loss_ctc=88.708, loss_att=69.286, acc=0.619, loss=75.132, grad_norm=3.807, loss_scale=1.000, learning_rate=1.612e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 04:27:43,677 (deepspeed_trainer:226) INFO: 7epoch:train:2401-2500batch: iter_time=1.072e-04, loss_ctc=99.213, loss_att=80.767, acc=0.611, loss=86.304, grad_norm=3.937, loss_scale=1.000, learning_rate=1.611e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 04:28:21,718 (deepspeed_trainer:226) INFO: 7epoch:train:2501-2600batch: iter_time=1.069e-04, loss_ctc=91.032, loss_att=76.022, acc=0.609, loss=80.499, grad_norm=3.691, loss_scale=1.000, learning_rate=1.610e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 04:28:59,726 (deepspeed_trainer:226) INFO: 7epoch:train:2601-2700batch: iter_time=1.053e-04, loss_ctc=107.084, loss_att=85.264, acc=0.591, loss=91.814, grad_norm=4.396, loss_scale=1.000, learning_rate=1.609e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 04:29:37,475 (deepspeed_trainer:226) INFO: 7epoch:train:2701-2800batch: iter_time=1.025e-04, loss_ctc=97.144, loss_att=82.249, acc=0.593, loss=86.672, grad_norm=4.139, loss_scale=1.000, learning_rate=1.609e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 04:30:16,128 (deepspeed_trainer:226) INFO: 7epoch:train:2801-2900batch: iter_time=1.071e-04, loss_ctc=106.637, loss_att=96.811, acc=0.588, loss=99.779, grad_norm=3.728, loss_scale=1.000, learning_rate=1.608e-04, step_time=0.386 [2024-11-29 04:30:53,273] [INFO] [logging.py:129:log_dist] [Rank 0] step=93000, skipped=0, lr=[np.float64(0.0001606430021107671)], mom=[[0.9, 0.98]] [2024-11-29 04:30:53,273] [INFO] [timer.py:264:stop] epoch=0/micro_step=93000/global_step=93000, RunningAvgSamplesPerSec=42.993070802941205, CurrSamplesPerSec=42.929693293801144, MemAllocated=2.04GB, MaxMemAllocated=15.53GB [cnode7-008:0/16] 2024-11-29 04:30:53,275 (deepspeed_trainer:226) INFO: 7epoch:train:2901-3000batch: iter_time=1.080e-04, loss_ctc=93.827, loss_att=82.184, acc=0.599, loss=85.637, grad_norm=3.537, loss_scale=1.000, learning_rate=1.607e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 04:31:30,852 (deepspeed_trainer:226) INFO: 7epoch:train:3001-3100batch: iter_time=1.103e-04, loss_ctc=117.779, loss_att=99.860, acc=0.586, loss=105.225, grad_norm=4.135, loss_scale=1.000, learning_rate=1.606e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 04:32:08,123 (deepspeed_trainer:226) INFO: 7epoch:train:3101-3200batch: iter_time=1.048e-04, loss_ctc=98.943, loss_att=85.340, acc=0.591, loss=89.429, grad_norm=4.072, loss_scale=1.000, learning_rate=1.605e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 04:32:44,826 (deepspeed_trainer:226) INFO: 7epoch:train:3201-3300batch: iter_time=1.064e-04, loss_ctc=99.511, loss_att=85.311, acc=0.601, loss=89.541, grad_norm=3.674, loss_scale=1.000, learning_rate=1.604e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 04:33:21,286 (deepspeed_trainer:226) INFO: 7epoch:train:3301-3400batch: iter_time=1.068e-04, loss_ctc=89.766, loss_att=76.538, acc=0.583, loss=80.538, grad_norm=3.431, loss_scale=1.000, learning_rate=1.603e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 04:33:59,006 (deepspeed_trainer:226) INFO: 7epoch:train:3401-3500batch: iter_time=1.062e-04, loss_ctc=97.992, loss_att=82.659, acc=0.601, loss=87.274, grad_norm=3.796, loss_scale=1.000, learning_rate=1.603e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 04:34:35,914 (deepspeed_trainer:226) INFO: 7epoch:train:3501-3600batch: iter_time=1.093e-04, loss_ctc=97.944, loss_att=89.142, acc=0.598, loss=91.765, grad_norm=3.644, loss_scale=1.000, learning_rate=1.602e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 04:35:12,938 (deepspeed_trainer:226) INFO: 7epoch:train:3601-3700batch: iter_time=1.104e-04, loss_ctc=105.125, loss_att=81.306, acc=0.622, loss=88.472, grad_norm=3.693, loss_scale=1.000, learning_rate=1.601e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 04:35:35,060 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-29 04:36:02,565 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 04:36:08,141 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 04:36:08,141 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-29 04:36:08,143 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 04:36:25,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:26,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:26,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:26,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:26,950] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:27,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:27,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:27,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,575] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:36:38,782] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:14,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:14,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:14,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:15,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:15,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:15,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:16,156] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:17,171] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:26,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:27,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:27,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:28,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:28,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:28,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:28,768] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:37:28,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:02,066] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:02,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:03,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:03,224] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:03,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:03,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:07,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:15,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:07,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:16,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:16,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:17,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:17,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:17,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:17,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:17,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:49,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:50,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:50,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:50,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:51,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:51,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:56,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:38:56,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:04,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:04,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:05,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:05,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:05,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:06,129] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:07,237] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:39:07,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 04:39:54,165 (deepspeed_trainer:226) INFO: 7epoch:train:3701-3800batch: iter_time=2.411, loss_ctc=91.451, loss_att=71.449, acc=0.616, loss=77.454, grad_norm=3.377, loss_scale=1.000, learning_rate=1.600e-04, step_time=0.401 [cnode7-008:0/16] 2024-11-29 04:40:32,080 (deepspeed_trainer:226) INFO: 7epoch:train:3801-3900batch: iter_time=1.023e-04, loss_ctc=101.837, loss_att=79.619, acc=0.629, loss=86.264, grad_norm=3.652, loss_scale=1.000, learning_rate=1.599e-04, step_time=0.379 [2024-11-29 04:41:10,099] [INFO] [logging.py:129:log_dist] [Rank 0] step=94000, skipped=0, lr=[np.float64(0.00015978624246283107)], mom=[[0.9, 0.98]] [2024-11-29 04:41:10,100] [INFO] [timer.py:264:stop] epoch=0/micro_step=94000/global_step=94000, RunningAvgSamplesPerSec=42.99763374135816, CurrSamplesPerSec=42.834555761556295, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 04:41:10,101 (deepspeed_trainer:226) INFO: 7epoch:train:3901-4000batch: iter_time=1.039e-04, loss_ctc=98.809, loss_att=93.377, acc=0.591, loss=94.953, grad_norm=3.786, loss_scale=1.000, learning_rate=1.598e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 04:41:48,427 (deepspeed_trainer:226) INFO: 7epoch:train:4001-4100batch: iter_time=1.063e-04, loss_ctc=94.957, loss_att=80.151, acc=0.585, loss=84.603, grad_norm=3.704, loss_scale=1.000, learning_rate=1.597e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 04:42:25,940 (deepspeed_trainer:226) INFO: 7epoch:train:4101-4200batch: iter_time=1.105e-04, loss_ctc=93.750, loss_att=75.846, acc=0.599, loss=81.199, grad_norm=3.819, loss_scale=1.000, learning_rate=1.597e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 04:43:02,754 (deepspeed_trainer:226) INFO: 7epoch:train:4201-4300batch: iter_time=1.071e-04, loss_ctc=93.075, loss_att=73.596, acc=0.612, loss=79.415, grad_norm=3.343, loss_scale=1.000, learning_rate=1.596e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 04:43:40,210 (deepspeed_trainer:226) INFO: 7epoch:train:4301-4400batch: iter_time=1.078e-04, loss_ctc=94.084, loss_att=72.691, acc=0.626, loss=79.121, grad_norm=3.812, loss_scale=1.000, learning_rate=1.595e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 04:44:16,974 (deepspeed_trainer:226) INFO: 7epoch:train:4401-4500batch: iter_time=1.069e-04, loss_ctc=102.554, loss_att=86.645, acc=0.586, loss=91.436, grad_norm=4.022, loss_scale=1.000, learning_rate=1.594e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 04:44:53,546 (deepspeed_trainer:226) INFO: 7epoch:train:4501-4600batch: iter_time=1.076e-04, loss_ctc=97.411, loss_att=73.579, acc=0.609, loss=80.778, grad_norm=3.701, loss_scale=1.000, learning_rate=1.593e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 04:45:30,279 (deepspeed_trainer:226) INFO: 7epoch:train:4601-4700batch: iter_time=1.068e-04, loss_ctc=96.413, loss_att=85.839, acc=0.594, loss=89.017, grad_norm=3.724, loss_scale=1.000, learning_rate=1.592e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 04:46:07,530 (deepspeed_trainer:226) INFO: 7epoch:train:4701-4800batch: iter_time=1.099e-04, loss_ctc=104.269, loss_att=90.146, acc=0.589, loss=94.426, grad_norm=3.973, loss_scale=1.000, learning_rate=1.592e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 04:46:45,614 (deepspeed_trainer:226) INFO: 7epoch:train:4801-4900batch: iter_time=1.093e-04, loss_ctc=97.061, loss_att=82.575, acc=0.597, loss=86.929, grad_norm=3.372, loss_scale=1.000, learning_rate=1.591e-04, step_time=0.380 [2024-11-29 04:47:24,118] [INFO] [logging.py:129:log_dist] [Rank 0] step=95000, skipped=0, lr=[np.float64(0.00015894304630765767)], mom=[[0.9, 0.98]] [2024-11-29 04:47:24,118] [INFO] [timer.py:264:stop] epoch=0/micro_step=95000/global_step=95000, RunningAvgSamplesPerSec=43.00424795178356, CurrSamplesPerSec=46.361679074256855, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 04:47:24,120 (deepspeed_trainer:226) INFO: 7epoch:train:4901-5000batch: iter_time=1.075e-04, loss_ctc=105.083, loss_att=89.274, acc=0.596, loss=94.006, grad_norm=3.554, loss_scale=1.000, learning_rate=1.590e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 04:48:02,122 (deepspeed_trainer:226) INFO: 7epoch:train:5001-5100batch: iter_time=1.086e-04, loss_ctc=104.108, loss_att=91.528, acc=0.581, loss=95.311, grad_norm=4.191, loss_scale=1.000, learning_rate=1.589e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 04:48:39,725 (deepspeed_trainer:226) INFO: 7epoch:train:5101-5200batch: iter_time=1.058e-04, loss_ctc=95.101, loss_att=77.606, acc=0.613, loss=82.843, grad_norm=3.430, loss_scale=1.000, learning_rate=1.588e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 04:49:16,862 (deepspeed_trainer:226) INFO: 7epoch:train:5201-5300batch: iter_time=1.030e-04, loss_ctc=89.976, loss_att=75.459, acc=0.584, loss=79.814, grad_norm=3.984, loss_scale=1.000, learning_rate=1.587e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 04:49:53,890 (deepspeed_trainer:226) INFO: 7epoch:train:5301-5400batch: iter_time=1.058e-04, loss_ctc=99.694, loss_att=85.460, acc=0.598, loss=89.741, grad_norm=3.521, loss_scale=1.000, learning_rate=1.587e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 04:50:30,532 (deepspeed_trainer:226) INFO: 7epoch:train:5401-5500batch: iter_time=1.093e-04, loss_ctc=90.702, loss_att=74.849, acc=0.614, loss=79.607, grad_norm=3.446, loss_scale=1.000, learning_rate=1.586e-04, step_time=0.366 [cnode7-008:0/16] 2024-11-29 04:51:07,500 (deepspeed_trainer:226) INFO: 7epoch:train:5501-5600batch: iter_time=1.111e-04, loss_ctc=103.559, loss_att=74.941, acc=0.617, loss=83.529, grad_norm=3.969, loss_scale=1.000, learning_rate=1.585e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 04:51:20,330 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-29 04:51:47,855 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 04:51:53,908 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 04:51:53,908 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-29 04:51:53,910 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 04:52:12,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:12,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:12,478] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:20,483] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:20,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:13,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:21,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:21,836] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:14,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:14,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:14,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:14,144] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:22,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:22,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:23,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:52:23,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:01,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:01,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:01,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:02,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:10,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:03,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:10,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:10,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:03,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:11,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:03,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:03,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:11,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:12,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:12,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:13,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:50,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:50,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:50,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:58,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:50,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:51,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:51,838] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:51,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:00,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:00,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:53:52,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:00,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:01,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:01,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:01,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:02,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:38,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:38,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:38,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:39,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:39,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:47,524] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:39,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:47,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:40,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:40,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:49,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:49,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:50,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:50,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:50,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 04:54:53,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 04:55:47,979 (deepspeed_trainer:226) INFO: 7epoch:train:5601-5700batch: iter_time=2.394, loss_ctc=99.416, loss_att=80.604, acc=0.626, loss=86.268, grad_norm=3.838, loss_scale=1.000, learning_rate=1.584e-04, step_time=0.411 [cnode7-008:0/16] 2024-11-29 04:56:26,287 (deepspeed_trainer:226) INFO: 7epoch:train:5701-5800batch: iter_time=1.063e-04, loss_ctc=94.726, loss_att=78.373, acc=0.627, loss=83.257, grad_norm=3.509, loss_scale=1.000, learning_rate=1.583e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 04:57:04,387 (deepspeed_trainer:226) INFO: 7epoch:train:5801-5900batch: iter_time=1.112e-04, loss_ctc=98.453, loss_att=92.890, acc=0.584, loss=94.547, grad_norm=4.027, loss_scale=1.000, learning_rate=1.582e-04, step_time=0.381 [2024-11-29 04:57:42,753] [INFO] [logging.py:129:log_dist] [Rank 0] step=96000, skipped=0, lr=[np.float64(0.00015811305950504524)], mom=[[0.9, 0.98]] [2024-11-29 04:57:42,754] [INFO] [timer.py:264:stop] epoch=0/micro_step=96000/global_step=96000, RunningAvgSamplesPerSec=43.004468709545954, CurrSamplesPerSec=42.21620007286112, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 04:57:42,755 (deepspeed_trainer:226) INFO: 7epoch:train:5901-6000batch: iter_time=1.105e-04, loss_ctc=88.779, loss_att=78.560, acc=0.602, loss=81.626, grad_norm=3.664, loss_scale=1.000, learning_rate=1.582e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 04:58:20,464 (deepspeed_trainer:226) INFO: 7epoch:train:6001-6100batch: iter_time=1.085e-04, loss_ctc=96.085, loss_att=75.808, acc=0.610, loss=81.868, grad_norm=3.932, loss_scale=1.000, learning_rate=1.581e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 04:58:58,472 (deepspeed_trainer:226) INFO: 7epoch:train:6101-6200batch: iter_time=1.086e-04, loss_ctc=92.348, loss_att=75.632, acc=0.620, loss=80.670, grad_norm=3.734, loss_scale=1.000, learning_rate=1.580e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 04:59:35,922 (deepspeed_trainer:226) INFO: 7epoch:train:6201-6300batch: iter_time=1.050e-04, loss_ctc=87.151, loss_att=67.574, acc=0.629, loss=73.434, grad_norm=3.667, loss_scale=1.000, learning_rate=1.579e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:00:13,721 (deepspeed_trainer:226) INFO: 7epoch:train:6301-6400batch: iter_time=1.054e-04, loss_ctc=108.517, loss_att=89.756, acc=0.597, loss=95.402, grad_norm=4.319, loss_scale=1.000, learning_rate=1.578e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:00:50,453 (deepspeed_trainer:226) INFO: 7epoch:train:6401-6500batch: iter_time=1.085e-04, loss_ctc=93.959, loss_att=72.522, acc=0.613, loss=78.950, grad_norm=4.006, loss_scale=1.000, learning_rate=1.577e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 05:01:27,597 (deepspeed_trainer:226) INFO: 7epoch:train:6501-6600batch: iter_time=1.083e-04, loss_ctc=104.487, loss_att=97.407, acc=0.588, loss=99.518, grad_norm=3.987, loss_scale=1.000, learning_rate=1.577e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 05:02:05,148 (deepspeed_trainer:226) INFO: 7epoch:train:6601-6700batch: iter_time=1.088e-04, loss_ctc=93.892, loss_att=82.547, acc=0.611, loss=85.959, grad_norm=4.008, loss_scale=1.000, learning_rate=1.576e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 05:02:42,312 (deepspeed_trainer:226) INFO: 7epoch:train:6701-6800batch: iter_time=1.106e-04, loss_ctc=106.652, loss_att=94.516, acc=0.584, loss=98.169, grad_norm=3.826, loss_scale=1.000, learning_rate=1.575e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 05:03:19,231 (deepspeed_trainer:226) INFO: 7epoch:train:6801-6900batch: iter_time=1.070e-04, loss_ctc=92.174, loss_att=74.725, acc=0.611, loss=79.945, grad_norm=3.809, loss_scale=1.000, learning_rate=1.574e-04, step_time=0.369 [2024-11-29 05:03:56,434] [INFO] [logging.py:129:log_dist] [Rank 0] step=97000, skipped=0, lr=[np.float64(0.0001572959407261944)], mom=[[0.9, 0.98]] [2024-11-29 05:03:56,434] [INFO] [timer.py:264:stop] epoch=0/micro_step=97000/global_step=97000, RunningAvgSamplesPerSec=43.01125693056889, CurrSamplesPerSec=46.51144679518109, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:03:56,436 (deepspeed_trainer:226) INFO: 7epoch:train:6901-7000batch: iter_time=1.101e-04, loss_ctc=105.294, loss_att=90.132, acc=0.597, loss=94.666, grad_norm=4.083, loss_scale=1.000, learning_rate=1.573e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 05:04:34,234 (deepspeed_trainer:226) INFO: 7epoch:train:7001-7100batch: iter_time=1.080e-04, loss_ctc=90.461, loss_att=75.129, acc=0.620, loss=79.716, grad_norm=3.527, loss_scale=1.000, learning_rate=1.573e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:05:11,703 (deepspeed_trainer:226) INFO: 7epoch:train:7101-7200batch: iter_time=1.097e-04, loss_ctc=88.765, loss_att=74.826, acc=0.588, loss=78.992, grad_norm=3.812, loss_scale=1.000, learning_rate=1.572e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:05:50,066 (deepspeed_trainer:226) INFO: 7epoch:train:7201-7300batch: iter_time=1.118e-04, loss_ctc=105.355, loss_att=97.308, acc=0.595, loss=99.693, grad_norm=3.633, loss_scale=1.000, learning_rate=1.571e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 05:06:27,993 (deepspeed_trainer:226) INFO: 7epoch:train:7301-7400batch: iter_time=1.076e-04, loss_ctc=85.952, loss_att=72.662, acc=0.627, loss=76.636, grad_norm=3.545, loss_scale=1.000, learning_rate=1.570e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:07:04,757 (deepspeed_trainer:226) INFO: 7epoch:train:7401-7500batch: iter_time=1.065e-04, loss_ctc=101.444, loss_att=72.909, acc=0.629, loss=81.457, grad_norm=4.020, loss_scale=1.000, learning_rate=1.569e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 05:07:08,662 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-29 05:07:36,131 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 05:07:42,593 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 05:07:42,593 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-29 05:07:42,596 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 05:08:08,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:08,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:09,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:09,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:01,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:09,941] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:01,921] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:10,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:10,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:10,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:02,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:58,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:59,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:51,360] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:59,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:59,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:51,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:51,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:51,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:52,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:52,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:52,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:08:52,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:00,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:00,591] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:00,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:02,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:46,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:46,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:39,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:47,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:39,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:48,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:40,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:48,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:40,643] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:40,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:49,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:41,153] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:41,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:51,588] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:52,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:09:45,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:26,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:35,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:35,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:27,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:36,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:28,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:36,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:28,730] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:29,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:29,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:37,686] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:29,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:39,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:40,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:40,375] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:10:35,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 05:11:46,151 (deepspeed_trainer:226) INFO: 7epoch:train:7501-7600batch: iter_time=2.427, loss_ctc=96.263, loss_att=75.496, acc=0.640, loss=81.737, grad_norm=4.502, loss_scale=1.000, learning_rate=1.569e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 05:12:24,199 (deepspeed_trainer:226) INFO: 7epoch:train:7601-7700batch: iter_time=1.027e-04, loss_ctc=93.620, loss_att=79.395, acc=0.621, loss=83.645, grad_norm=3.685, loss_scale=1.000, learning_rate=1.568e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 05:13:01,594 (deepspeed_trainer:226) INFO: 7epoch:train:7701-7800batch: iter_time=1.042e-04, loss_ctc=96.161, loss_att=88.062, acc=0.588, loss=90.472, grad_norm=3.805, loss_scale=1.000, learning_rate=1.567e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:13:38,977 (deepspeed_trainer:226) INFO: 7epoch:train:7801-7900batch: iter_time=1.050e-04, loss_ctc=90.279, loss_att=81.427, acc=0.607, loss=84.046, grad_norm=3.407, loss_scale=1.000, learning_rate=1.566e-04, step_time=0.373 [2024-11-29 05:14:16,196] [INFO] [logging.py:129:log_dist] [Rank 0] step=98000, skipped=0, lr=[np.float64(0.00015649136086391658)], mom=[[0.9, 0.98]] [2024-11-29 05:14:16,197] [INFO] [timer.py:264:stop] epoch=0/micro_step=98000/global_step=98000, RunningAvgSamplesPerSec=43.01378481195885, CurrSamplesPerSec=43.59390878546842, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:14:16,200 (deepspeed_trainer:226) INFO: 7epoch:train:7901-8000batch: iter_time=1.064e-04, loss_ctc=93.988, loss_att=73.047, acc=0.617, loss=79.344, grad_norm=3.653, loss_scale=1.000, learning_rate=1.565e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 05:14:55,148 (deepspeed_trainer:226) INFO: 7epoch:train:8001-8100batch: iter_time=1.061e-04, loss_ctc=88.343, loss_att=70.862, acc=0.625, loss=76.104, grad_norm=3.339, loss_scale=1.000, learning_rate=1.565e-04, step_time=0.389 [cnode7-008:0/16] 2024-11-29 05:15:33,992 (deepspeed_trainer:226) INFO: 7epoch:train:8101-8200batch: iter_time=1.064e-04, loss_ctc=96.442, loss_att=78.707, acc=0.622, loss=84.051, grad_norm=3.862, loss_scale=1.000, learning_rate=1.564e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 05:16:11,911 (deepspeed_trainer:226) INFO: 7epoch:train:8201-8300batch: iter_time=1.054e-04, loss_ctc=100.993, loss_att=82.517, acc=0.591, loss=88.097, grad_norm=3.752, loss_scale=1.000, learning_rate=1.563e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:16:50,319 (deepspeed_trainer:226) INFO: 7epoch:train:8301-8400batch: iter_time=1.056e-04, loss_ctc=91.692, loss_att=68.117, acc=0.629, loss=75.193, grad_norm=3.648, loss_scale=1.000, learning_rate=1.562e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 05:17:27,880 (deepspeed_trainer:226) INFO: 7epoch:train:8401-8500batch: iter_time=1.072e-04, loss_ctc=104.118, loss_att=96.678, acc=0.598, loss=98.945, grad_norm=3.772, loss_scale=1.000, learning_rate=1.561e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 05:18:05,542 (deepspeed_trainer:226) INFO: 7epoch:train:8501-8600batch: iter_time=1.116e-04, loss_ctc=87.932, loss_att=79.133, acc=0.605, loss=81.752, grad_norm=3.656, loss_scale=1.000, learning_rate=1.561e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 05:18:43,504 (deepspeed_trainer:226) INFO: 7epoch:train:8601-8700batch: iter_time=1.126e-04, loss_ctc=110.292, loss_att=93.529, acc=0.606, loss=98.573, grad_norm=3.913, loss_scale=1.000, learning_rate=1.560e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:19:21,045 (deepspeed_trainer:226) INFO: 7epoch:train:8701-8800batch: iter_time=1.089e-04, loss_ctc=101.567, loss_att=81.518, acc=0.594, loss=87.537, grad_norm=3.951, loss_scale=1.000, learning_rate=1.559e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 05:19:58,236 (deepspeed_trainer:226) INFO: 7epoch:train:8801-8900batch: iter_time=1.126e-04, loss_ctc=95.060, loss_att=84.125, acc=0.622, loss=87.430, grad_norm=3.502, loss_scale=1.000, learning_rate=1.558e-04, step_time=0.372 [2024-11-29 05:20:35,457] [INFO] [logging.py:129:log_dist] [Rank 0] step=99000, skipped=0, lr=[np.float64(0.000155699002475692)], mom=[[0.9, 0.98]] [2024-11-29 05:20:35,457] [INFO] [timer.py:264:stop] epoch=0/micro_step=99000/global_step=99000, RunningAvgSamplesPerSec=43.013514098586754, CurrSamplesPerSec=46.474437217463326, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:20:35,458 (deepspeed_trainer:226) INFO: 7epoch:train:8901-9000batch: iter_time=1.121e-04, loss_ctc=90.200, loss_att=73.282, acc=0.613, loss=78.353, grad_norm=3.485, loss_scale=1.000, learning_rate=1.557e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 05:21:12,226 (deepspeed_trainer:226) INFO: 7epoch:train:9001-9100batch: iter_time=1.111e-04, loss_ctc=91.633, loss_att=76.651, acc=0.600, loss=81.169, grad_norm=3.451, loss_scale=1.000, learning_rate=1.557e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 05:21:49,412 (deepspeed_trainer:226) INFO: 7epoch:train:9101-9200batch: iter_time=1.127e-04, loss_ctc=96.155, loss_att=85.968, acc=0.604, loss=89.009, grad_norm=3.475, loss_scale=1.000, learning_rate=1.556e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 05:22:26,667 (deepspeed_trainer:226) INFO: 7epoch:train:9201-9300batch: iter_time=1.120e-04, loss_ctc=102.138, loss_att=80.854, acc=0.627, loss=87.223, grad_norm=3.725, loss_scale=1.000, learning_rate=1.555e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 05:22:58,546 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-29 05:23:26,471 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 05:23:31,983 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 05:23:31,983 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-29 05:23:31,986 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 05:23:49,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:50,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:51,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:51,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:51,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:52,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:52,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:23:52,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:00,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:00,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:00,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:01,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:01,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:01,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:02,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:02,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:39,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:39,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:39,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:40,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:40,755] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:41,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:41,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:50,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:50,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:51,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:51,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:51,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:52,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:52,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:44,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:24:54,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:27,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:27,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:27,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:28,066] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:28,839] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:29,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:29,686] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:40,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:40,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:40,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:41,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:41,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:41,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:41,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:35,524] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:25:44,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:15,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:15,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:15,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:16,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:17,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:18,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:18,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,885] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:30,970] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:31,975] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:24,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:26:33,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 05:27:09,307 (deepspeed_trainer:226) INFO: 7epoch:train:9301-9400batch: iter_time=2.438, loss_ctc=89.034, loss_att=68.989, acc=0.636, loss=75.021, grad_norm=3.539, loss_scale=1.000, learning_rate=1.554e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 05:27:46,585 (deepspeed_trainer:226) INFO: 7epoch:train:9401-9500batch: iter_time=1.161e-04, loss_ctc=92.664, loss_att=70.485, acc=0.628, loss=77.149, grad_norm=3.797, loss_scale=1.000, learning_rate=1.553e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 05:28:24,743 (deepspeed_trainer:226) INFO: 7epoch:train:9501-9600batch: iter_time=1.087e-04, loss_ctc=101.700, loss_att=91.487, acc=0.609, loss=94.531, grad_norm=3.802, loss_scale=1.000, learning_rate=1.553e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 05:29:02,102 (deepspeed_trainer:226) INFO: 7epoch:train:9601-9700batch: iter_time=1.101e-04, loss_ctc=94.421, loss_att=83.911, acc=0.597, loss=87.065, grad_norm=3.715, loss_scale=1.000, learning_rate=1.552e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:29:40,006 (deepspeed_trainer:226) INFO: 7epoch:train:9701-9800batch: iter_time=1.065e-04, loss_ctc=93.975, loss_att=78.022, acc=0.593, loss=82.810, grad_norm=4.024, loss_scale=1.000, learning_rate=1.551e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:30:16,901 (deepspeed_trainer:226) INFO: 7epoch:train:9801-9900batch: iter_time=1.093e-04, loss_ctc=84.623, loss_att=65.517, acc=0.629, loss=71.253, grad_norm=3.320, loss_scale=1.000, learning_rate=1.550e-04, step_time=0.368 [2024-11-29 05:30:54,387] [INFO] [logging.py:129:log_dist] [Rank 0] step=100000, skipped=0, lr=[np.float64(0.00015491855925743687)], mom=[[0.9, 0.98]] [2024-11-29 05:30:54,388] [INFO] [timer.py:264:stop] epoch=0/micro_step=100000/global_step=100000, RunningAvgSamplesPerSec=43.01815107947554, CurrSamplesPerSec=42.57427833283581, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:30:54,389 (deepspeed_trainer:226) INFO: 7epoch:train:9901-10000batch: iter_time=1.103e-04, loss_ctc=95.809, loss_att=75.934, acc=0.622, loss=81.845, grad_norm=3.892, loss_scale=1.000, learning_rate=1.550e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 05:31:31,784 (deepspeed_trainer:226) INFO: 7epoch:train:10001-10100batch: iter_time=1.107e-04, loss_ctc=86.478, loss_att=72.281, acc=0.622, loss=76.521, grad_norm=3.551, loss_scale=1.000, learning_rate=1.549e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:32:08,771 (deepspeed_trainer:226) INFO: 7epoch:train:10101-10200batch: iter_time=1.162e-04, loss_ctc=102.310, loss_att=81.737, acc=0.600, loss=87.954, grad_norm=3.931, loss_scale=1.000, learning_rate=1.548e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 05:32:45,642 (deepspeed_trainer:226) INFO: 7epoch:train:10201-10300batch: iter_time=1.139e-04, loss_ctc=93.141, loss_att=76.847, acc=0.607, loss=81.782, grad_norm=3.914, loss_scale=1.000, learning_rate=1.547e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 05:33:23,545 (deepspeed_trainer:226) INFO: 7epoch:train:10301-10400batch: iter_time=1.120e-04, loss_ctc=104.012, loss_att=91.510, acc=0.598, loss=95.237, grad_norm=3.896, loss_scale=1.000, learning_rate=1.546e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:34:01,183 (deepspeed_trainer:226) INFO: 7epoch:train:10401-10500batch: iter_time=1.174e-04, loss_ctc=89.910, loss_att=77.657, acc=0.607, loss=81.321, grad_norm=3.745, loss_scale=1.000, learning_rate=1.546e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 05:34:38,971 (deepspeed_trainer:226) INFO: 7epoch:train:10501-10600batch: iter_time=1.112e-04, loss_ctc=112.896, loss_att=94.171, acc=0.599, loss=99.798, grad_norm=3.956, loss_scale=1.000, learning_rate=1.545e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:35:17,351 (deepspeed_trainer:226) INFO: 7epoch:train:10601-10700batch: iter_time=1.099e-04, loss_ctc=94.228, loss_att=81.896, acc=0.600, loss=85.594, grad_norm=3.905, loss_scale=1.000, learning_rate=1.544e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 05:35:54,854 (deepspeed_trainer:226) INFO: 7epoch:train:10701-10800batch: iter_time=1.098e-04, loss_ctc=95.420, loss_att=80.115, acc=0.612, loss=84.659, grad_norm=3.989, loss_scale=1.000, learning_rate=1.543e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:36:33,402 (deepspeed_trainer:226) INFO: 7epoch:train:10801-10900batch: iter_time=1.082e-04, loss_ctc=85.651, loss_att=72.161, acc=0.593, loss=76.199, grad_norm=3.970, loss_scale=1.000, learning_rate=1.543e-04, step_time=0.386 [2024-11-29 05:37:11,737] [INFO] [logging.py:129:log_dist] [Rank 0] step=101000, skipped=0, lr=[np.float64(0.00015414973554599858)], mom=[[0.9, 0.98]] [2024-11-29 05:37:11,738] [INFO] [timer.py:264:stop] epoch=0/micro_step=101000/global_step=101000, RunningAvgSamplesPerSec=43.0203908690043, CurrSamplesPerSec=45.16775839563477, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:37:11,739 (deepspeed_trainer:226) INFO: 7epoch:train:10901-11000batch: iter_time=1.122e-04, loss_ctc=93.274, loss_att=75.896, acc=0.614, loss=81.101, grad_norm=3.744, loss_scale=1.000, learning_rate=1.542e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 05:37:48,911 (deepspeed_trainer:226) INFO: 7epoch:train:11001-11100batch: iter_time=1.092e-04, loss_ctc=95.294, loss_att=81.863, acc=0.606, loss=85.881, grad_norm=3.795, loss_scale=1.000, learning_rate=1.541e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 05:38:26,684 (deepspeed_trainer:226) INFO: 7epoch:train:11101-11200batch: iter_time=1.125e-04, loss_ctc=101.396, loss_att=77.103, acc=0.628, loss=84.408, grad_norm=3.890, loss_scale=1.000, learning_rate=1.540e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:38:49,400 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-29 05:39:17,400 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 05:39:22,893 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 05:39:22,893 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-29 05:39:22,895 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 05:39:49,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:40,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:41,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:50,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:50,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:41,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:42,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:50,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:51,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:51,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:42,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:51,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:42,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:42,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:42,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:39:52,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:28,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:29,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:29,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:30,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:39,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:39,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:30,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:39,604] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:31,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:31,755] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:31,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:40,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:40,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:40,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:41,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:40:41,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:16,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:17,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:17,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:18,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:26,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:27,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:19,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:27,784] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:19,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:20,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:20,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:29,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:30,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:30,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:30,503] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:41:34,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:03,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:04,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:13,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:05,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:05,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:14,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:14,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:06,714] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:06,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:16,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:08,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:17,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:08,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:17,569] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:18,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:42:22,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 05:43:07,288 (deepspeed_trainer:226) INFO: 7epoch:train:11201-11300batch: iter_time=2.431, loss_ctc=87.593, loss_att=67.381, acc=0.629, loss=73.464, grad_norm=3.946, loss_scale=1.000, learning_rate=1.540e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 05:43:45,652 (deepspeed_trainer:226) INFO: 7epoch:train:11301-11400batch: iter_time=1.011e-04, loss_ctc=98.989, loss_att=75.894, acc=0.641, loss=82.812, grad_norm=3.767, loss_scale=1.000, learning_rate=1.539e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 05:44:23,085 (deepspeed_trainer:226) INFO: 7epoch:train:11401-11500batch: iter_time=1.085e-04, loss_ctc=95.501, loss_att=88.507, acc=0.603, loss=90.612, grad_norm=3.849, loss_scale=1.000, learning_rate=1.538e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:45:00,978 (deepspeed_trainer:226) INFO: 7epoch:train:11501-11600batch: iter_time=1.067e-04, loss_ctc=91.319, loss_att=75.928, acc=0.600, loss=80.556, grad_norm=3.581, loss_scale=1.000, learning_rate=1.537e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:45:38,695 (deepspeed_trainer:226) INFO: 7epoch:train:11601-11700batch: iter_time=1.114e-04, loss_ctc=90.705, loss_att=72.497, acc=0.610, loss=77.942, grad_norm=4.150, loss_scale=1.000, learning_rate=1.537e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 05:46:16,545 (deepspeed_trainer:226) INFO: 7epoch:train:11701-11800batch: iter_time=1.083e-04, loss_ctc=89.548, loss_att=69.872, acc=0.624, loss=75.769, grad_norm=3.382, loss_scale=1.000, learning_rate=1.536e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:46:54,566 (deepspeed_trainer:226) INFO: 7epoch:train:11801-11900batch: iter_time=1.065e-04, loss_ctc=90.045, loss_att=69.586, acc=0.640, loss=75.715, grad_norm=4.167, loss_scale=1.000, learning_rate=1.535e-04, step_time=0.380 [2024-11-29 05:47:31,919] [INFO] [logging.py:129:log_dist] [Rank 0] step=102000, skipped=0, lr=[np.float64(0.00015339224584854326)], mom=[[0.9, 0.98]] [2024-11-29 05:47:31,920] [INFO] [timer.py:264:stop] epoch=0/micro_step=102000/global_step=102000, RunningAvgSamplesPerSec=43.02268376986508, CurrSamplesPerSec=45.88906369503279, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:47:31,921 (deepspeed_trainer:226) INFO: 7epoch:train:11901-12000batch: iter_time=1.080e-04, loss_ctc=98.026, loss_att=82.910, acc=0.600, loss=87.454, grad_norm=4.108, loss_scale=1.000, learning_rate=1.534e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 05:48:09,169 (deepspeed_trainer:226) INFO: 7epoch:train:12001-12100batch: iter_time=1.067e-04, loss_ctc=93.637, loss_att=69.891, acc=0.620, loss=76.986, grad_norm=4.123, loss_scale=1.000, learning_rate=1.534e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 05:48:46,502 (deepspeed_trainer:226) INFO: 7epoch:train:12101-12200batch: iter_time=1.094e-04, loss_ctc=92.541, loss_att=82.215, acc=0.607, loss=85.296, grad_norm=3.565, loss_scale=1.000, learning_rate=1.533e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 05:49:23,275 (deepspeed_trainer:226) INFO: 7epoch:train:12201-12300batch: iter_time=1.078e-04, loss_ctc=98.927, loss_att=84.455, acc=0.606, loss=88.802, grad_norm=3.885, loss_scale=1.000, learning_rate=1.532e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 05:50:00,777 (deepspeed_trainer:226) INFO: 7epoch:train:12301-12400batch: iter_time=1.069e-04, loss_ctc=94.418, loss_att=79.861, acc=0.607, loss=84.226, grad_norm=4.288, loss_scale=1.000, learning_rate=1.531e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 05:50:38,654 (deepspeed_trainer:226) INFO: 7epoch:train:12401-12500batch: iter_time=1.263e-04, loss_ctc=102.148, loss_att=85.154, acc=0.609, loss=90.255, grad_norm=4.203, loss_scale=1.000, learning_rate=1.531e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 05:51:16,470 (deepspeed_trainer:226) INFO: 7epoch:train:12501-12600batch: iter_time=1.075e-04, loss_ctc=100.261, loss_att=87.200, acc=0.596, loss=91.099, grad_norm=4.575, loss_scale=1.000, learning_rate=1.530e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:51:53,604 (deepspeed_trainer:226) INFO: 7epoch:train:12601-12700batch: iter_time=1.121e-04, loss_ctc=91.889, loss_att=74.646, acc=0.624, loss=79.831, grad_norm=3.868, loss_scale=1.000, learning_rate=1.529e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 05:52:31,469 (deepspeed_trainer:226) INFO: 7epoch:train:12701-12800batch: iter_time=1.419e-04, loss_ctc=84.981, loss_att=71.534, acc=0.598, loss=75.546, grad_norm=3.959, loss_scale=1.000, learning_rate=1.528e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:53:09,039 (deepspeed_trainer:226) INFO: 7epoch:train:12801-12900batch: iter_time=1.102e-04, loss_ctc=96.763, loss_att=82.073, acc=0.610, loss=86.486, grad_norm=3.866, loss_scale=1.000, learning_rate=1.528e-04, step_time=0.375 [2024-11-29 05:53:46,901] [INFO] [logging.py:129:log_dist] [Rank 0] step=103000, skipped=0, lr=[np.float64(0.00015264581439713483)], mom=[[0.9, 0.98]] [2024-11-29 05:53:46,902] [INFO] [timer.py:264:stop] epoch=0/micro_step=103000/global_step=103000, RunningAvgSamplesPerSec=43.02772839566594, CurrSamplesPerSec=42.96864256921193, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 05:53:46,903 (deepspeed_trainer:226) INFO: 7epoch:train:12901-13000batch: iter_time=1.100e-04, loss_ctc=87.777, loss_att=71.835, acc=0.625, loss=76.648, grad_norm=3.744, loss_scale=1.000, learning_rate=1.527e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 05:54:24,205 (deepspeed_trainer:226) INFO: 7epoch:train:13001-13100batch: iter_time=1.096e-04, loss_ctc=99.314, loss_att=71.279, acc=0.630, loss=79.698, grad_norm=3.973, loss_scale=1.000, learning_rate=1.526e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 05:54:37,711 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-29 05:55:05,563 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 05:55:11,119 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 05:55:11,119 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-29 05:55:11,121 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 05:55:29,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:29,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:29,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:37,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:30,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:30,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:38,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:38,594] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:31,150] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:31,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:31,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:39,129] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:39,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:39,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:40,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:55:40,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:17,729] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:18,877] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:19,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:19,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:27,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:19,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:19,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:20,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:28,103] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:28,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:29,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:29,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:22,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:30,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:31,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:56:32,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:05,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:07,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:07,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:07,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:16,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:09,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:16,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:09,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:09,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:17,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:18,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:18,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:19,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:13,477] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:22,412] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:24,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:54,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:55,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:56,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:56,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:04,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:05,650] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:05,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:58,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:06,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:57:59,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:07,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:07,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:00,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:09,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:02,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 05:58:14,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 05:59:09,231 (deepspeed_trainer:226) INFO: 7epoch:train:13101-13200batch: iter_time=2.425, loss_ctc=95.169, loss_att=78.144, acc=0.635, loss=83.257, grad_norm=4.205, loss_scale=1.000, learning_rate=1.525e-04, step_time=0.425 [cnode7-008:0/16] 2024-11-29 05:59:47,063 (deepspeed_trainer:226) INFO: 7epoch:train:13201-13300batch: iter_time=1.074e-04, loss_ctc=92.136, loss_att=75.499, acc=0.637, loss=80.477, grad_norm=4.023, loss_scale=1.000, learning_rate=1.525e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 06:00:24,547 (deepspeed_trainer:226) INFO: 7epoch:train:13301-13400batch: iter_time=1.069e-04, loss_ctc=95.917, loss_att=89.274, acc=0.594, loss=91.256, grad_norm=4.236, loss_scale=1.000, learning_rate=1.524e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 06:01:02,252 (deepspeed_trainer:226) INFO: 7epoch:train:13401-13500batch: iter_time=1.031e-04, loss_ctc=86.034, loss_att=77.278, acc=0.608, loss=79.939, grad_norm=3.661, loss_scale=1.000, learning_rate=1.523e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 06:01:40,091 (deepspeed_trainer:226) INFO: 7epoch:train:13501-13600batch: iter_time=1.050e-04, loss_ctc=93.628, loss_att=74.075, acc=0.619, loss=79.930, grad_norm=3.996, loss_scale=1.000, learning_rate=1.522e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 06:02:17,157 (deepspeed_trainer:226) INFO: 7epoch:train:13601-13700batch: iter_time=1.087e-04, loss_ctc=88.869, loss_att=71.919, acc=0.630, loss=76.986, grad_norm=3.660, loss_scale=1.000, learning_rate=1.522e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:02:54,152 (deepspeed_trainer:226) INFO: 7epoch:train:13701-13800batch: iter_time=1.105e-04, loss_ctc=84.013, loss_att=65.441, acc=0.640, loss=71.026, grad_norm=3.482, loss_scale=1.000, learning_rate=1.521e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 06:03:31,764 (deepspeed_trainer:226) INFO: 7epoch:train:13801-13900batch: iter_time=1.117e-04, loss_ctc=105.423, loss_att=87.916, acc=0.606, loss=93.145, grad_norm=4.315, loss_scale=1.000, learning_rate=1.520e-04, step_time=0.376 [2024-11-29 06:04:08,817] [INFO] [logging.py:129:log_dist] [Rank 0] step=104000, skipped=0, lr=[np.float64(0.00015191017472692713)], mom=[[0.9, 0.98]] [2024-11-29 06:04:08,817] [INFO] [timer.py:264:stop] epoch=0/micro_step=104000/global_step=104000, RunningAvgSamplesPerSec=43.02740637870971, CurrSamplesPerSec=44.078297577531096, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:04:08,818 (deepspeed_trainer:226) INFO: 7epoch:train:13901-14000batch: iter_time=1.110e-04, loss_ctc=91.150, loss_att=69.147, acc=0.622, loss=75.769, grad_norm=3.924, loss_scale=1.000, learning_rate=1.519e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:04:47,370 (deepspeed_trainer:226) INFO: 7epoch:train:14001-14100batch: iter_time=1.127e-04, loss_ctc=102.234, loss_att=94.763, acc=0.600, loss=97.005, grad_norm=3.730, loss_scale=1.000, learning_rate=1.519e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 06:05:25,235 (deepspeed_trainer:226) INFO: 7epoch:train:14101-14200batch: iter_time=1.101e-04, loss_ctc=89.093, loss_att=78.699, acc=0.622, loss=81.858, grad_norm=3.764, loss_scale=1.000, learning_rate=1.518e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 06:06:03,394 (deepspeed_trainer:226) INFO: 7epoch:train:14201-14300batch: iter_time=1.076e-04, loss_ctc=103.526, loss_att=90.771, acc=0.599, loss=94.566, grad_norm=3.916, loss_scale=1.000, learning_rate=1.517e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 06:06:41,055 (deepspeed_trainer:226) INFO: 7epoch:train:14301-14400batch: iter_time=1.347e-04, loss_ctc=89.827, loss_att=71.933, acc=0.621, loss=77.310, grad_norm=3.899, loss_scale=1.000, learning_rate=1.517e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 06:07:20,757 (deepspeed_trainer:226) INFO: 7epoch:train:14401-14500batch: iter_time=1.462e-04, loss_ctc=102.870, loss_att=86.932, acc=0.611, loss=91.723, grad_norm=4.212, loss_scale=1.000, learning_rate=1.516e-04, step_time=0.397 [cnode7-008:0/16] 2024-11-29 06:07:58,698 (deepspeed_trainer:226) INFO: 7epoch:train:14501-14600batch: iter_time=1.166e-04, loss_ctc=87.903, loss_att=70.979, acc=0.636, loss=76.089, grad_norm=3.581, loss_scale=1.000, learning_rate=1.515e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 06:08:36,234 (deepspeed_trainer:226) INFO: 7epoch:train:14601-14700batch: iter_time=1.077e-04, loss_ctc=83.997, loss_att=71.547, acc=0.601, loss=75.285, grad_norm=3.663, loss_scale=1.000, learning_rate=1.514e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 06:09:13,529 (deepspeed_trainer:226) INFO: 7epoch:train:14701-14800batch: iter_time=1.096e-04, loss_ctc=103.514, loss_att=93.802, acc=0.609, loss=96.756, grad_norm=3.716, loss_scale=1.000, learning_rate=1.514e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 06:09:50,415 (deepspeed_trainer:226) INFO: 7epoch:train:14801-14900batch: iter_time=1.080e-04, loss_ctc=82.540, loss_att=68.812, acc=0.641, loss=72.918, grad_norm=3.303, loss_scale=1.000, learning_rate=1.513e-04, step_time=0.369 [2024-11-29 06:10:27,882] [INFO] [logging.py:129:log_dist] [Rank 0] step=105000, skipped=0, lr=[np.float64(0.0001511850692765037)], mom=[[0.9, 0.98]] [2024-11-29 06:10:27,882] [INFO] [timer.py:264:stop] epoch=0/micro_step=105000/global_step=105000, RunningAvgSamplesPerSec=43.027990746723304, CurrSamplesPerSec=45.70678851650177, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:10:27,884 (deepspeed_trainer:226) INFO: 7epoch:train:14901-15000batch: iter_time=1.039e-04, loss_ctc=99.272, loss_att=71.097, acc=0.639, loss=79.558, grad_norm=4.030, loss_scale=1.000, learning_rate=1.512e-04, step_time=0.374 [2024-11-29 06:10:33,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:42,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:34,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:34,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:35,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:43,859] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:35,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:35,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:35,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:35,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:44,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:48,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:49,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:58,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:50,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:50,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:50,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:59,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:59,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:51,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:51,307] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:00,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:10:51,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:01,023] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:01,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:01,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:01,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:03,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:14,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:05,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:05,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:06,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:06,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:06,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:07,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:07,103] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:16,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:16,133] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:16,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:16,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:17,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:17,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:17,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:18,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:29,650] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:21,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:21,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:22,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:31,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:22,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:22,885] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:23,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:23,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:32,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:32,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:32,756] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:33,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:33,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:33,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:11:46,203] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 7 is about to be saved! [2024-11-29 06:11:46,236] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/mp_rank_00_model_states.pt [2024-11-29 06:11:46,236] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/mp_rank_00_model_states.pt... [2024-11-29 06:11:48,305] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/mp_rank_00_model_states.pt. [2024-11-29 06:11:48,462] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,702] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,702] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,703] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,463] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,704] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,464] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,468] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,469] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,709] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,469] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,710] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,710] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,470] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-29 06:11:39,711] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-29 06:11:48,471] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-29 06:11:49,042] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,042] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,042] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,050] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,054] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,054] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,056] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,057] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,057] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,340] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,341] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,341] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,115] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,116] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,116] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,372] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,372] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,372] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,446] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,446] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,446] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,206] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,206] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,206] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,215] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,215] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,215] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,455] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,456] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,456] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,218] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,218] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-29 06:11:49,218] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,218] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-29 06:11:49,218] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:49,218] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,465] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,465] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,465] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,482] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,482] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,482] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,541] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,541] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,541] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [2024-11-29 06:11:40,545] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-29 06:11:40,545] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_7/7/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-29 06:11:40,545] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 7 is ready now! [cnode7-008:0/16] 2024-11-29 06:11:49,310 (deepspeed_trainer:158) INFO: 7epoch results: [train] iter_time=0.129, loss_ctc=96.602, loss_att=80.573, acc=0.607, loss=85.383, grad_norm=3.798, loss_scale=1.000, learning_rate=1.570e-04, step_time=0.376, time=2 hours, 6 minutes and 32.98 seconds, total_count=105007, gpu_max_cached_mem_GB=28.338, [valid] loss_ctc=10.625, cer_ctc=0.259, loss_att=11.125, acc=0.783, cer=0.135, wer=1.000, loss=11.000, time=1 minute and 9.35 seconds, total_count=7, gpu_max_cached_mem_GB=28.338 [cnode7-008:0/16] 2024-11-29 06:11:51,863 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-29 06:12:18,255 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 06:12:23,917 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 06:12:23,917 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-29 06:12:23,920 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 06:12:35,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:35,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:36,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:45,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:36,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:36,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:37,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:45,784] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:37,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:37,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:46,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:46,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:47,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:47,164] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:47,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:12:48,894] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:22,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:22,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:23,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:23,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:24,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:25,103] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:25,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:25,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:34,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:35,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:36,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:36,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:37,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:38,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:38,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:13:39,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:09,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:10,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:11,218] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:11,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:11,900] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:12,329] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:12,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:12,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:22,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:23,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:25,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:25,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:26,318] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:27,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:27,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:28,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:56,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:57,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:58,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:58,529] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:59,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:14:59,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:00,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:00,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:11,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:12,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:13,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:15,362] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:15,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:16,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:17,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:15:17,872] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 06:16:21,799 (deepspeed_trainer:226) INFO: 8epoch:train:1-100batch: iter_time=2.307, loss_ctc=86.909, loss_att=74.461, acc=0.605, loss=78.203, grad_norm=3.523, loss_scale=1.000, learning_rate=1.511e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 06:16:58,736 (deepspeed_trainer:226) INFO: 8epoch:train:101-200batch: iter_time=1.061e-04, loss_ctc=81.616, loss_att=63.300, acc=0.622, loss=68.814, grad_norm=3.602, loss_scale=1.000, learning_rate=1.511e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 06:17:36,724 (deepspeed_trainer:226) INFO: 8epoch:train:201-300batch: iter_time=1.053e-04, loss_ctc=92.234, loss_att=78.501, acc=0.606, loss=82.612, grad_norm=3.651, loss_scale=1.000, learning_rate=1.510e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 06:18:14,872 (deepspeed_trainer:226) INFO: 8epoch:train:301-400batch: iter_time=1.036e-04, loss_ctc=89.493, loss_att=81.481, acc=0.601, loss=83.913, grad_norm=3.911, loss_scale=1.000, learning_rate=1.509e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 06:18:52,535 (deepspeed_trainer:226) INFO: 8epoch:train:401-500batch: iter_time=1.061e-04, loss_ctc=101.506, loss_att=78.379, acc=0.608, loss=85.353, grad_norm=4.126, loss_scale=1.000, learning_rate=1.509e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 06:19:30,753 (deepspeed_trainer:226) INFO: 8epoch:train:501-600batch: iter_time=1.094e-04, loss_ctc=100.105, loss_att=84.699, acc=0.595, loss=89.310, grad_norm=4.087, loss_scale=1.000, learning_rate=1.508e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:20:08,583 (deepspeed_trainer:226) INFO: 8epoch:train:601-700batch: iter_time=1.140e-04, loss_ctc=90.182, loss_att=75.481, acc=0.602, loss=79.915, grad_norm=3.507, loss_scale=1.000, learning_rate=1.507e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 06:20:46,233 (deepspeed_trainer:226) INFO: 8epoch:train:701-800batch: iter_time=1.135e-04, loss_ctc=92.111, loss_att=77.157, acc=0.616, loss=81.629, grad_norm=3.520, loss_scale=1.000, learning_rate=1.506e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 06:21:24,634 (deepspeed_trainer:226) INFO: 8epoch:train:801-900batch: iter_time=1.129e-04, loss_ctc=92.977, loss_att=73.677, acc=0.610, loss=79.470, grad_norm=3.927, loss_scale=1.000, learning_rate=1.506e-04, step_time=0.384 [2024-11-29 06:22:01,983] [INFO] [logging.py:129:log_dist] [Rank 0] step=106000, skipped=0, lr=[np.float64(0.0001504702490090042)], mom=[[0.9, 0.98]] [2024-11-29 06:22:01,984] [INFO] [timer.py:264:stop] epoch=0/micro_step=106000/global_step=106000, RunningAvgSamplesPerSec=43.0274404680836, CurrSamplesPerSec=45.22376421839706, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:22:01,985 (deepspeed_trainer:226) INFO: 8epoch:train:901-1000batch: iter_time=1.028e-04, loss_ctc=101.703, loss_att=83.419, acc=0.593, loss=88.861, grad_norm=4.522, loss_scale=1.000, learning_rate=1.505e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 06:22:39,231 (deepspeed_trainer:226) INFO: 8epoch:train:1001-1100batch: iter_time=1.039e-04, loss_ctc=88.927, loss_att=72.657, acc=0.611, loss=77.523, grad_norm=3.594, loss_scale=1.000, learning_rate=1.504e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 06:23:16,719 (deepspeed_trainer:226) INFO: 8epoch:train:1101-1200batch: iter_time=1.043e-04, loss_ctc=102.222, loss_att=80.461, acc=0.598, loss=86.989, grad_norm=4.426, loss_scale=1.000, learning_rate=1.504e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 06:23:53,796 (deepspeed_trainer:226) INFO: 8epoch:train:1201-1300batch: iter_time=1.076e-04, loss_ctc=95.406, loss_att=72.756, acc=0.613, loss=79.565, grad_norm=3.745, loss_scale=1.000, learning_rate=1.503e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:24:31,841 (deepspeed_trainer:226) INFO: 8epoch:train:1301-1400batch: iter_time=1.041e-04, loss_ctc=95.641, loss_att=74.706, acc=0.622, loss=80.988, grad_norm=4.038, loss_scale=1.000, learning_rate=1.502e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 06:25:10,132 (deepspeed_trainer:226) INFO: 8epoch:train:1401-1500batch: iter_time=1.041e-04, loss_ctc=109.359, loss_att=82.778, acc=0.615, loss=90.724, grad_norm=4.453, loss_scale=1.000, learning_rate=1.502e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:25:47,824 (deepspeed_trainer:226) INFO: 8epoch:train:1501-1600batch: iter_time=1.045e-04, loss_ctc=90.916, loss_att=70.444, acc=0.630, loss=76.638, grad_norm=3.753, loss_scale=1.000, learning_rate=1.501e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 06:26:24,279 (deepspeed_trainer:226) INFO: 8epoch:train:1601-1700batch: iter_time=1.049e-04, loss_ctc=74.099, loss_att=57.571, acc=0.620, loss=62.556, grad_norm=3.322, loss_scale=1.000, learning_rate=1.500e-04, step_time=0.364 [cnode7-008:0/16] 2024-11-29 06:27:02,485 (deepspeed_trainer:226) INFO: 8epoch:train:1701-1800batch: iter_time=1.058e-04, loss_ctc=102.373, loss_att=77.996, acc=0.620, loss=85.308, grad_norm=4.028, loss_scale=1.000, learning_rate=1.499e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:27:34,911 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-29 06:28:02,747 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 06:28:08,440 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 06:28:08,440 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-29 06:28:08,442 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 06:28:36,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:36,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:36,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:36,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:37,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:37,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:28,478] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:37,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:28,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:28,883] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:38,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:28,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:29,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:29,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:29,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:28:29,153] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:27,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:27,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:27,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:28,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:28,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:28,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:28,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:29,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:19,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:19,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:20,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:20,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:20,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:20,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:21,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:29:21,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:07,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:07,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:17,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:08,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:08,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:17,815] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:08,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:08,710] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:18,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:09,108] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:18,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:18,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:19,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:19,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:11,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:22,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:55,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:05,609] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:56,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:56,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:56,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:56,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:56,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:06,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:30:57,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:06,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:07,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:08,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:08,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:09,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:00,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:31:12,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 06:31:50,443 (deepspeed_trainer:226) INFO: 8epoch:train:1801-1900batch: iter_time=2.465, loss_ctc=99.646, loss_att=89.500, acc=0.596, loss=92.552, grad_norm=3.976, loss_scale=1.000, learning_rate=1.499e-04, step_time=0.414 [2024-11-29 06:32:27,388] [INFO] [logging.py:129:log_dist] [Rank 0] step=107000, skipped=0, lr=[np.float64(0.00014976547305277138)], mom=[[0.9, 0.98]] [2024-11-29 06:32:27,391] [INFO] [timer.py:264:stop] epoch=0/micro_step=107000/global_step=107000, RunningAvgSamplesPerSec=43.027612559228274, CurrSamplesPerSec=44.73467393517456, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:32:27,392 (deepspeed_trainer:226) INFO: 8epoch:train:1901-2000batch: iter_time=1.091e-04, loss_ctc=84.441, loss_att=64.011, acc=0.624, loss=70.167, grad_norm=3.602, loss_scale=1.000, learning_rate=1.498e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 06:33:04,468 (deepspeed_trainer:226) INFO: 8epoch:train:2001-2100batch: iter_time=1.090e-04, loss_ctc=78.432, loss_att=64.066, acc=0.618, loss=68.371, grad_norm=3.458, loss_scale=1.000, learning_rate=1.497e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:33:42,268 (deepspeed_trainer:226) INFO: 8epoch:train:2101-2200batch: iter_time=1.157e-04, loss_ctc=86.532, loss_att=72.694, acc=0.624, loss=76.869, grad_norm=3.429, loss_scale=1.000, learning_rate=1.497e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 06:34:19,183 (deepspeed_trainer:226) INFO: 8epoch:train:2201-2300batch: iter_time=1.131e-04, loss_ctc=96.976, loss_att=83.454, acc=0.605, loss=87.508, grad_norm=3.784, loss_scale=1.000, learning_rate=1.496e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 06:34:56,570 (deepspeed_trainer:226) INFO: 8epoch:train:2301-2400batch: iter_time=1.132e-04, loss_ctc=92.113, loss_att=68.590, acc=0.616, loss=75.636, grad_norm=3.643, loss_scale=1.000, learning_rate=1.495e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 06:35:34,160 (deepspeed_trainer:226) INFO: 8epoch:train:2401-2500batch: iter_time=1.080e-04, loss_ctc=99.362, loss_att=85.524, acc=0.597, loss=89.648, grad_norm=3.923, loss_scale=1.000, learning_rate=1.495e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 06:36:11,565 (deepspeed_trainer:226) INFO: 8epoch:train:2501-2600batch: iter_time=1.161e-04, loss_ctc=93.416, loss_att=77.378, acc=0.610, loss=82.177, grad_norm=3.445, loss_scale=1.000, learning_rate=1.494e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:36:48,965 (deepspeed_trainer:226) INFO: 8epoch:train:2601-2700batch: iter_time=1.127e-04, loss_ctc=87.183, loss_att=71.319, acc=0.618, loss=76.061, grad_norm=3.414, loss_scale=1.000, learning_rate=1.493e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 06:37:25,959 (deepspeed_trainer:226) INFO: 8epoch:train:2701-2800batch: iter_time=1.154e-04, loss_ctc=93.954, loss_att=71.934, acc=0.616, loss=78.521, grad_norm=4.200, loss_scale=1.000, learning_rate=1.492e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:38:03,176 (deepspeed_trainer:226) INFO: 8epoch:train:2801-2900batch: iter_time=1.083e-04, loss_ctc=101.320, loss_att=89.160, acc=0.587, loss=92.774, grad_norm=3.967, loss_scale=1.000, learning_rate=1.492e-04, step_time=0.372 [2024-11-29 06:38:40,214] [INFO] [logging.py:129:log_dist] [Rank 0] step=108000, skipped=0, lr=[np.float64(0.0001490705083603411)], mom=[[0.9, 0.98]] [2024-11-29 06:38:40,214] [INFO] [timer.py:264:stop] epoch=0/micro_step=108000/global_step=108000, RunningAvgSamplesPerSec=43.03429299404152, CurrSamplesPerSec=46.33187958168073, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:38:40,216 (deepspeed_trainer:226) INFO: 8epoch:train:2901-3000batch: iter_time=1.080e-04, loss_ctc=85.355, loss_att=70.491, acc=0.616, loss=74.968, grad_norm=3.643, loss_scale=1.000, learning_rate=1.491e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:39:17,649 (deepspeed_trainer:226) INFO: 8epoch:train:3001-3100batch: iter_time=1.073e-04, loss_ctc=95.059, loss_att=71.504, acc=0.615, loss=78.561, grad_norm=3.806, loss_scale=1.000, learning_rate=1.490e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 06:39:54,140 (deepspeed_trainer:226) INFO: 8epoch:train:3101-3200batch: iter_time=1.075e-04, loss_ctc=88.513, loss_att=67.997, acc=0.624, loss=74.142, grad_norm=3.439, loss_scale=1.000, learning_rate=1.490e-04, step_time=0.365 [cnode7-008:0/16] 2024-11-29 06:40:31,493 (deepspeed_trainer:226) INFO: 8epoch:train:3201-3300batch: iter_time=1.056e-04, loss_ctc=98.963, loss_att=79.747, acc=0.624, loss=85.471, grad_norm=3.856, loss_scale=1.000, learning_rate=1.489e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 06:41:08,727 (deepspeed_trainer:226) INFO: 8epoch:train:3301-3400batch: iter_time=1.046e-04, loss_ctc=106.279, loss_att=75.809, acc=0.627, loss=84.931, grad_norm=4.299, loss_scale=1.000, learning_rate=1.488e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 06:41:45,950 (deepspeed_trainer:226) INFO: 8epoch:train:3401-3500batch: iter_time=1.080e-04, loss_ctc=86.051, loss_att=67.581, acc=0.627, loss=73.126, grad_norm=3.729, loss_scale=1.000, learning_rate=1.488e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 06:42:22,081 (deepspeed_trainer:226) INFO: 8epoch:train:3501-3600batch: iter_time=1.098e-04, loss_ctc=83.064, loss_att=62.435, acc=0.624, loss=68.622, grad_norm=3.224, loss_scale=1.000, learning_rate=1.487e-04, step_time=0.361 [cnode7-008:0/16] 2024-11-29 06:42:59,577 (deepspeed_trainer:226) INFO: 8epoch:train:3601-3700batch: iter_time=1.067e-04, loss_ctc=96.823, loss_att=83.540, acc=0.615, loss=87.501, grad_norm=3.934, loss_scale=1.000, learning_rate=1.486e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 06:43:22,280 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-29 06:43:49,261 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 06:43:55,175 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 06:43:55,175 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-29 06:43:55,178 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 06:44:11,475] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:11,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:12,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:22,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:22,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:12,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:13,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:13,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:13,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:23,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:13,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:23,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:24,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:24,108] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:24,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:24,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:44:59,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:00,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:00,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:01,369] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:01,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:01,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:01,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:11,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:12,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:13,143] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:13,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:13,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:14,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:14,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:14,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:04,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:47,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:48,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:49,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:49,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:49,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:50,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:50,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:01,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:01,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:01,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:02,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:02,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:02,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:04,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:45:54,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:05,412] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:36,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:36,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:36,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:37,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:37,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:49,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:50,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:50,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:40,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:50,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:51,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:42,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:52,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:44,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:54,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 06:46:56,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 06:47:42,603 (deepspeed_trainer:226) INFO: 8epoch:train:3701-3800batch: iter_time=2.429, loss_ctc=100.097, loss_att=86.948, acc=0.610, loss=90.944, grad_norm=3.821, loss_scale=1.000, learning_rate=1.486e-04, step_time=0.401 [cnode7-008:0/16] 2024-11-29 06:48:20,144 (deepspeed_trainer:226) INFO: 8epoch:train:3801-3900batch: iter_time=1.074e-04, loss_ctc=78.514, loss_att=58.804, acc=0.640, loss=64.719, grad_norm=3.166, loss_scale=1.000, learning_rate=1.485e-04, step_time=0.375 [2024-11-29 06:48:58,190] [INFO] [logging.py:129:log_dist] [Rank 0] step=109000, skipped=0, lr=[np.float64(0.00014838512938467944)], mom=[[0.9, 0.98]] [2024-11-29 06:48:58,190] [INFO] [timer.py:264:stop] epoch=0/micro_step=109000/global_step=109000, RunningAvgSamplesPerSec=43.038619275491286, CurrSamplesPerSec=44.43209266521404, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:48:58,192 (deepspeed_trainer:226) INFO: 8epoch:train:3901-4000batch: iter_time=1.109e-04, loss_ctc=82.357, loss_att=69.374, acc=0.610, loss=73.266, grad_norm=3.698, loss_scale=1.000, learning_rate=1.484e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 06:49:35,655 (deepspeed_trainer:226) INFO: 8epoch:train:4001-4100batch: iter_time=1.069e-04, loss_ctc=84.958, loss_att=75.927, acc=0.623, loss=78.682, grad_norm=3.294, loss_scale=1.000, learning_rate=1.484e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 06:50:13,037 (deepspeed_trainer:226) INFO: 8epoch:train:4101-4200batch: iter_time=1.103e-04, loss_ctc=95.324, loss_att=78.752, acc=0.621, loss=83.712, grad_norm=3.727, loss_scale=1.000, learning_rate=1.483e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 06:50:51,117 (deepspeed_trainer:226) INFO: 8epoch:train:4201-4300batch: iter_time=1.040e-04, loss_ctc=94.787, loss_att=77.313, acc=0.615, loss=82.567, grad_norm=3.696, loss_scale=1.000, learning_rate=1.482e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 06:51:29,328 (deepspeed_trainer:226) INFO: 8epoch:train:4301-4400batch: iter_time=1.044e-04, loss_ctc=94.433, loss_att=81.976, acc=0.605, loss=85.706, grad_norm=3.760, loss_scale=1.000, learning_rate=1.481e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:52:06,631 (deepspeed_trainer:226) INFO: 8epoch:train:4401-4500batch: iter_time=1.087e-04, loss_ctc=85.959, loss_att=69.703, acc=0.629, loss=74.557, grad_norm=3.198, loss_scale=1.000, learning_rate=1.481e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 06:52:44,756 (deepspeed_trainer:226) INFO: 8epoch:train:4501-4600batch: iter_time=1.078e-04, loss_ctc=91.257, loss_att=79.533, acc=0.618, loss=83.061, grad_norm=3.653, loss_scale=1.000, learning_rate=1.480e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 06:53:22,474 (deepspeed_trainer:226) INFO: 8epoch:train:4601-4700batch: iter_time=1.080e-04, loss_ctc=96.521, loss_att=77.855, acc=0.607, loss=83.478, grad_norm=4.145, loss_scale=1.000, learning_rate=1.479e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 06:54:00,037 (deepspeed_trainer:226) INFO: 8epoch:train:4701-4800batch: iter_time=1.068e-04, loss_ctc=92.797, loss_att=78.585, acc=0.610, loss=82.817, grad_norm=3.612, loss_scale=1.000, learning_rate=1.479e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 06:54:37,891 (deepspeed_trainer:226) INFO: 8epoch:train:4801-4900batch: iter_time=1.082e-04, loss_ctc=92.838, loss_att=75.721, acc=0.618, loss=80.882, grad_norm=3.952, loss_scale=1.000, learning_rate=1.478e-04, step_time=0.379 [2024-11-29 06:55:16,390] [INFO] [logging.py:129:log_dist] [Rank 0] step=110000, skipped=0, lr=[np.float64(0.00014770911777164442)], mom=[[0.9, 0.98]] [2024-11-29 06:55:16,390] [INFO] [timer.py:264:stop] epoch=0/micro_step=110000/global_step=110000, RunningAvgSamplesPerSec=43.03950876453801, CurrSamplesPerSec=42.15115621497503, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 06:55:16,392 (deepspeed_trainer:226) INFO: 8epoch:train:4901-5000batch: iter_time=1.105e-04, loss_ctc=94.488, loss_att=73.641, acc=0.621, loss=79.894, grad_norm=3.685, loss_scale=1.000, learning_rate=1.477e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 06:55:54,860 (deepspeed_trainer:226) INFO: 8epoch:train:5001-5100batch: iter_time=1.088e-04, loss_ctc=88.110, loss_att=76.080, acc=0.623, loss=79.729, grad_norm=3.354, loss_scale=1.000, learning_rate=1.477e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 06:56:32,629 (deepspeed_trainer:226) INFO: 8epoch:train:5101-5200batch: iter_time=1.086e-04, loss_ctc=98.532, loss_att=77.472, acc=0.623, loss=83.773, grad_norm=4.001, loss_scale=1.000, learning_rate=1.476e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 06:57:10,833 (deepspeed_trainer:226) INFO: 8epoch:train:5201-5300batch: iter_time=1.093e-04, loss_ctc=106.592, loss_att=76.679, acc=0.635, loss=85.635, grad_norm=4.081, loss_scale=1.000, learning_rate=1.475e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:57:47,839 (deepspeed_trainer:226) INFO: 8epoch:train:5301-5400batch: iter_time=1.083e-04, loss_ctc=74.009, loss_att=59.194, acc=0.639, loss=63.633, grad_norm=3.257, loss_scale=1.000, learning_rate=1.475e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 06:58:25,207 (deepspeed_trainer:226) INFO: 8epoch:train:5401-5500batch: iter_time=1.108e-04, loss_ctc=93.854, loss_att=73.241, acc=0.627, loss=79.442, grad_norm=3.871, loss_scale=1.000, learning_rate=1.474e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 06:59:03,398 (deepspeed_trainer:226) INFO: 8epoch:train:5501-5600batch: iter_time=1.123e-04, loss_ctc=91.515, loss_att=77.421, acc=0.629, loss=81.638, grad_norm=3.901, loss_scale=1.000, learning_rate=1.473e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 06:59:17,002 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-29 06:59:44,405 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 06:59:50,222 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 06:59:50,222 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-29 06:59:50,225 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 07:00:07,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:07,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:18,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:18,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:18,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:09,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:18,971] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:19,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:19,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:19,921] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:20,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:55,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:56,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:56,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:57,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:57,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:06,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:57,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:00:57,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:07,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:08,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:08,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:08,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:08,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:09,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:01,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:11,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:43,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:44,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:45,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:54,704] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:45,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:45,866] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:45,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:46,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:56,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:56,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:57,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:58,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:58,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:58,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:01:50,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:00,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:31,448] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:32,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:33,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:33,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:33,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:42,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:34,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:34,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:45,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:46,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:46,398] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:46,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:48,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:39,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:49,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:02:50,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 07:03:46,286 (deepspeed_trainer:226) INFO: 8epoch:train:5601-5700batch: iter_time=2.418, loss_ctc=91.365, loss_att=78.927, acc=0.609, loss=82.670, grad_norm=3.531, loss_scale=1.000, learning_rate=1.473e-04, step_time=0.410 [cnode7-008:0/16] 2024-11-29 07:04:24,772 (deepspeed_trainer:226) INFO: 8epoch:train:5701-5800batch: iter_time=1.111e-04, loss_ctc=79.171, loss_att=56.382, acc=0.644, loss=63.219, grad_norm=3.319, loss_scale=1.000, learning_rate=1.472e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 07:05:03,459 (deepspeed_trainer:226) INFO: 8epoch:train:5801-5900batch: iter_time=1.145e-04, loss_ctc=84.972, loss_att=76.023, acc=0.603, loss=78.716, grad_norm=3.774, loss_scale=1.000, learning_rate=1.471e-04, step_time=0.387 [2024-11-29 07:05:40,955] [INFO] [logging.py:129:log_dist] [Rank 0] step=111000, skipped=0, lr=[np.float64(0.00014704226206772148)], mom=[[0.9, 0.98]] [2024-11-29 07:05:40,956] [INFO] [timer.py:264:stop] epoch=0/micro_step=111000/global_step=111000, RunningAvgSamplesPerSec=43.03577333417713, CurrSamplesPerSec=46.004703419256025, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 07:05:40,957 (deepspeed_trainer:226) INFO: 8epoch:train:5901-6000batch: iter_time=1.129e-04, loss_ctc=86.567, loss_att=75.512, acc=0.627, loss=78.838, grad_norm=3.472, loss_scale=1.000, learning_rate=1.471e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 07:06:19,752 (deepspeed_trainer:226) INFO: 8epoch:train:6001-6100batch: iter_time=1.143e-04, loss_ctc=90.536, loss_att=71.856, acc=0.622, loss=77.438, grad_norm=3.809, loss_scale=1.000, learning_rate=1.470e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 07:06:58,302 (deepspeed_trainer:226) INFO: 8epoch:train:6101-6200batch: iter_time=1.116e-04, loss_ctc=93.467, loss_att=78.909, acc=0.617, loss=83.280, grad_norm=3.936, loss_scale=1.000, learning_rate=1.469e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 07:07:35,765 (deepspeed_trainer:226) INFO: 8epoch:train:6201-6300batch: iter_time=1.114e-04, loss_ctc=93.122, loss_att=77.937, acc=0.600, loss=82.502, grad_norm=3.670, loss_scale=1.000, learning_rate=1.469e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:08:13,119 (deepspeed_trainer:226) INFO: 8epoch:train:6301-6400batch: iter_time=1.111e-04, loss_ctc=90.117, loss_att=71.471, acc=0.633, loss=77.053, grad_norm=3.368, loss_scale=1.000, learning_rate=1.468e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 07:08:51,045 (deepspeed_trainer:226) INFO: 8epoch:train:6401-6500batch: iter_time=1.097e-04, loss_ctc=87.426, loss_att=73.529, acc=0.613, loss=77.690, grad_norm=3.551, loss_scale=1.000, learning_rate=1.467e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 07:09:28,283 (deepspeed_trainer:226) INFO: 8epoch:train:6501-6600batch: iter_time=1.110e-04, loss_ctc=96.294, loss_att=77.079, acc=0.604, loss=82.816, grad_norm=4.375, loss_scale=1.000, learning_rate=1.467e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:10:06,518 (deepspeed_trainer:226) INFO: 8epoch:train:6601-6700batch: iter_time=1.136e-04, loss_ctc=87.388, loss_att=72.112, acc=0.621, loss=76.733, grad_norm=3.500, loss_scale=1.000, learning_rate=1.466e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 07:10:43,633 (deepspeed_trainer:226) INFO: 8epoch:train:6701-6800batch: iter_time=1.145e-04, loss_ctc=96.088, loss_att=78.848, acc=0.615, loss=84.051, grad_norm=4.144, loss_scale=1.000, learning_rate=1.465e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 07:11:21,868 (deepspeed_trainer:226) INFO: 8epoch:train:6801-6900batch: iter_time=1.183e-04, loss_ctc=90.199, loss_att=68.528, acc=0.616, loss=75.015, grad_norm=3.763, loss_scale=1.000, learning_rate=1.465e-04, step_time=0.382 [2024-11-29 07:12:00,528] [INFO] [logging.py:129:log_dist] [Rank 0] step=112000, skipped=0, lr=[np.float64(0.00014638435744214297)], mom=[[0.9, 0.98]] [2024-11-29 07:12:00,528] [INFO] [timer.py:264:stop] epoch=0/micro_step=112000/global_step=112000, RunningAvgSamplesPerSec=43.03540261799937, CurrSamplesPerSec=43.61983571238591, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 07:12:00,530 (deepspeed_trainer:226) INFO: 8epoch:train:6901-7000batch: iter_time=1.168e-04, loss_ctc=88.745, loss_att=72.159, acc=0.626, loss=77.145, grad_norm=3.655, loss_scale=1.000, learning_rate=1.464e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 07:12:38,864 (deepspeed_trainer:226) INFO: 8epoch:train:7001-7100batch: iter_time=1.168e-04, loss_ctc=99.857, loss_att=73.790, acc=0.630, loss=81.629, grad_norm=3.842, loss_scale=1.000, learning_rate=1.464e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 07:13:17,537 (deepspeed_trainer:226) INFO: 8epoch:train:7101-7200batch: iter_time=1.240e-04, loss_ctc=99.974, loss_att=73.524, acc=0.644, loss=81.431, grad_norm=3.972, loss_scale=1.000, learning_rate=1.463e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 07:13:54,312 (deepspeed_trainer:226) INFO: 8epoch:train:7201-7300batch: iter_time=1.165e-04, loss_ctc=74.251, loss_att=58.431, acc=0.636, loss=63.152, grad_norm=3.143, loss_scale=1.000, learning_rate=1.462e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 07:14:32,760 (deepspeed_trainer:226) INFO: 8epoch:train:7301-7400batch: iter_time=1.166e-04, loss_ctc=94.741, loss_att=73.130, acc=0.624, loss=79.609, grad_norm=3.916, loss_scale=1.000, learning_rate=1.462e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 07:15:09,884 (deepspeed_trainer:226) INFO: 8epoch:train:7401-7500batch: iter_time=1.151e-04, loss_ctc=92.846, loss_att=77.805, acc=0.620, loss=82.326, grad_norm=3.684, loss_scale=1.000, learning_rate=1.461e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 07:15:13,687 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-29 07:15:40,873 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 07:15:46,664 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 07:15:46,664 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-29 07:15:46,667 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 07:16:03,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:03,816] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:04,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:14,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:04,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:14,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:05,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:05,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:05,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:05,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:15,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:15,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:15,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:16,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:16,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:16,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:51,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:52,330] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:52,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:02,885] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:53,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:03,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:54,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:54,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:54,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:04,085] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:16:55,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:04,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:05,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:05,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:05,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:06,478] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:41,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:41,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:41,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:42,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:51,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:42,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:42,759] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:43,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:52,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:53,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:53,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:54,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:54,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:54,816] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:45,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:17:55,900] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:28,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:29,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:29,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:30,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:30,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:40,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:31,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:42,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:42,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:43,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:33,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:43,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:43,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:35,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:45,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:18:45,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 07:19:50,687 (deepspeed_trainer:226) INFO: 8epoch:train:7501-7600batch: iter_time=2.427, loss_ctc=83.707, loss_att=73.699, acc=0.622, loss=76.695, grad_norm=3.441, loss_scale=1.000, learning_rate=1.460e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 07:20:29,464 (deepspeed_trainer:226) INFO: 8epoch:train:7601-7700batch: iter_time=1.140e-04, loss_ctc=78.263, loss_att=60.198, acc=0.639, loss=65.621, grad_norm=3.454, loss_scale=1.000, learning_rate=1.460e-04, step_time=0.387 [cnode7-008:0/16] 2024-11-29 07:21:06,990 (deepspeed_trainer:226) INFO: 8epoch:train:7701-7800batch: iter_time=1.228e-04, loss_ctc=87.204, loss_att=75.444, acc=0.622, loss=78.956, grad_norm=3.627, loss_scale=1.000, learning_rate=1.459e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 07:21:45,024 (deepspeed_trainer:226) INFO: 8epoch:train:7801-7900batch: iter_time=1.088e-04, loss_ctc=85.109, loss_att=77.052, acc=0.627, loss=79.457, grad_norm=3.552, loss_scale=1.000, learning_rate=1.458e-04, step_time=0.380 [2024-11-29 07:22:23,531] [INFO] [logging.py:129:log_dist] [Rank 0] step=113000, skipped=0, lr=[np.float64(0.00014573520542256302)], mom=[[0.9, 0.98]] [2024-11-29 07:22:23,531] [INFO] [timer.py:264:stop] epoch=0/micro_step=113000/global_step=113000, RunningAvgSamplesPerSec=43.034391337901916, CurrSamplesPerSec=43.92849271480107, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 07:22:23,533 (deepspeed_trainer:226) INFO: 8epoch:train:7901-8000batch: iter_time=1.096e-04, loss_ctc=96.647, loss_att=74.236, acc=0.625, loss=80.947, grad_norm=3.695, loss_scale=1.000, learning_rate=1.458e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 07:23:01,138 (deepspeed_trainer:226) INFO: 8epoch:train:8001-8100batch: iter_time=1.102e-04, loss_ctc=94.628, loss_att=80.855, acc=0.614, loss=85.027, grad_norm=4.176, loss_scale=1.000, learning_rate=1.457e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 07:23:39,707 (deepspeed_trainer:226) INFO: 8epoch:train:8101-8200batch: iter_time=1.081e-04, loss_ctc=87.273, loss_att=76.231, acc=0.614, loss=79.562, grad_norm=3.607, loss_scale=1.000, learning_rate=1.456e-04, step_time=0.385 [cnode7-008:0/16] 2024-11-29 07:24:17,302 (deepspeed_trainer:226) INFO: 8epoch:train:8201-8300batch: iter_time=1.086e-04, loss_ctc=88.349, loss_att=74.262, acc=0.636, loss=78.481, grad_norm=3.587, loss_scale=1.000, learning_rate=1.456e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 07:24:55,912 (deepspeed_trainer:226) INFO: 8epoch:train:8301-8400batch: iter_time=1.123e-04, loss_ctc=88.946, loss_att=70.260, acc=0.628, loss=75.872, grad_norm=4.065, loss_scale=1.000, learning_rate=1.455e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 07:25:33,631 (deepspeed_trainer:226) INFO: 8epoch:train:8401-8500batch: iter_time=1.122e-04, loss_ctc=94.256, loss_att=81.709, acc=0.603, loss=85.461, grad_norm=4.403, loss_scale=1.000, learning_rate=1.454e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 07:26:11,783 (deepspeed_trainer:226) INFO: 8epoch:train:8501-8600batch: iter_time=1.137e-04, loss_ctc=85.835, loss_att=69.983, acc=0.628, loss=74.715, grad_norm=3.549, loss_scale=1.000, learning_rate=1.454e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 07:26:49,015 (deepspeed_trainer:226) INFO: 8epoch:train:8601-8700batch: iter_time=1.077e-04, loss_ctc=93.788, loss_att=76.919, acc=0.615, loss=82.011, grad_norm=3.907, loss_scale=1.000, learning_rate=1.453e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:27:26,264 (deepspeed_trainer:226) INFO: 8epoch:train:8701-8800batch: iter_time=1.098e-04, loss_ctc=91.877, loss_att=72.597, acc=0.622, loss=78.386, grad_norm=3.699, loss_scale=1.000, learning_rate=1.453e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:28:03,605 (deepspeed_trainer:226) INFO: 8epoch:train:8801-8900batch: iter_time=1.163e-04, loss_ctc=90.754, loss_att=72.274, acc=0.637, loss=77.827, grad_norm=3.909, loss_scale=1.000, learning_rate=1.452e-04, step_time=0.373 [2024-11-29 07:28:41,754] [INFO] [logging.py:129:log_dist] [Rank 0] step=114000, skipped=0, lr=[np.float64(0.0001450946136435135)], mom=[[0.9, 0.98]] [2024-11-29 07:28:41,755] [INFO] [timer.py:264:stop] epoch=0/micro_step=114000/global_step=114000, RunningAvgSamplesPerSec=43.035528221102076, CurrSamplesPerSec=37.83784234807003, MemAllocated=2.04GB, MaxMemAllocated=15.54GB [cnode7-008:0/16] 2024-11-29 07:28:41,756 (deepspeed_trainer:226) INFO: 8epoch:train:8901-9000batch: iter_time=1.217e-04, loss_ctc=104.618, loss_att=81.157, acc=0.631, loss=88.207, grad_norm=4.379, loss_scale=1.000, learning_rate=1.451e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 07:29:20,111 (deepspeed_trainer:226) INFO: 8epoch:train:9001-9100batch: iter_time=1.209e-04, loss_ctc=87.463, loss_att=67.228, acc=0.646, loss=73.291, grad_norm=3.774, loss_scale=1.000, learning_rate=1.451e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 07:29:57,510 (deepspeed_trainer:226) INFO: 8epoch:train:9101-9200batch: iter_time=1.145e-04, loss_ctc=70.601, loss_att=54.928, acc=0.637, loss=59.638, grad_norm=3.207, loss_scale=1.000, learning_rate=1.450e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:30:35,671 (deepspeed_trainer:226) INFO: 8epoch:train:9201-9300batch: iter_time=1.160e-04, loss_ctc=98.202, loss_att=75.151, acc=0.638, loss=82.099, grad_norm=4.178, loss_scale=1.000, learning_rate=1.449e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 07:31:08,507 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-29 07:31:35,088 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 07:31:40,885 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.0", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.0", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.0", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.0", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 07:31:40,885 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, [cnode7-008:0/16] 2024-11-29 07:31:40,888 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 07:31:57,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:57,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:57,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:58,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:08,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:58,314] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:58,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:09,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:59,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:31:59,488] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:09,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:10,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:10,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:10,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:11,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:11,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:45,573] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:45,852] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:46,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:46,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:46,813] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:47,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:57,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:58,078] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:58,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:48,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:48,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:58,609] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:59,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:32:59,673] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:00,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:03,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:35,335] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:35,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:35,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:35,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:36,245] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:36,254] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:36,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:36,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:48,412] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:48,589] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:48,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:48,759] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:49,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:49,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:50,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:33:56,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:23,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:24,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:24,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:24,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:25,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:25,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:25,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:36,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:37,412] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:37,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:27,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:38,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:38,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:38,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:38,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:34:47,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 07:35:23,376 (deepspeed_trainer:226) INFO: 8epoch:train:9301-9400batch: iter_time=2.408, loss_ctc=95.640, loss_att=89.495, acc=0.611, loss=91.359, grad_norm=3.745, loss_scale=1.000, learning_rate=1.449e-04, step_time=0.469 [cnode7-008:0/16] 2024-11-29 07:36:00,579 (deepspeed_trainer:226) INFO: 8epoch:train:9401-9500batch: iter_time=1.119e-04, loss_ctc=81.494, loss_att=61.505, acc=0.640, loss=67.504, grad_norm=3.490, loss_scale=1.000, learning_rate=1.448e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:36:38,890 (deepspeed_trainer:226) INFO: 8epoch:train:9501-9600batch: iter_time=1.058e-04, loss_ctc=75.824, loss_att=61.689, acc=0.635, loss=65.899, grad_norm=3.631, loss_scale=1.000, learning_rate=1.447e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 07:37:16,875 (deepspeed_trainer:226) INFO: 8epoch:train:9601-9700batch: iter_time=1.154e-04, loss_ctc=84.492, loss_att=70.811, acc=0.636, loss=74.913, grad_norm=3.295, loss_scale=1.000, learning_rate=1.447e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 07:37:54,902 (deepspeed_trainer:226) INFO: 8epoch:train:9701-9800batch: iter_time=1.080e-04, loss_ctc=92.958, loss_att=80.108, acc=0.628, loss=83.993, grad_norm=3.623, loss_scale=1.000, learning_rate=1.446e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 07:38:32,435 (deepspeed_trainer:226) INFO: 8epoch:train:9801-9900batch: iter_time=1.124e-04, loss_ctc=87.764, loss_att=65.307, acc=0.633, loss=72.081, grad_norm=3.565, loss_scale=1.000, learning_rate=1.446e-04, step_time=0.375 [2024-11-29 07:39:10,293] [INFO] [logging.py:129:log_dist] [Rank 0] step=115000, skipped=0, lr=[np.float64(0.00014446239560691627)], mom=[[0.9, 0.98]] [2024-11-29 07:39:10,293] [INFO] [timer.py:264:stop] epoch=0/micro_step=115000/global_step=115000, RunningAvgSamplesPerSec=43.02704114470607, CurrSamplesPerSec=42.98289855341018, MemAllocated=2.04GB, MaxMemAllocated=16.57GB [cnode7-008:0/16] 2024-11-29 07:39:10,295 (deepspeed_trainer:226) INFO: 8epoch:train:9901-10000batch: iter_time=1.100e-04, loss_ctc=95.746, loss_att=82.130, acc=0.613, loss=86.258, grad_norm=4.222, loss_scale=1.000, learning_rate=1.445e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 07:39:48,641 (deepspeed_trainer:226) INFO: 8epoch:train:10001-10100batch: iter_time=1.112e-04, loss_ctc=91.116, loss_att=77.901, acc=0.623, loss=81.841, grad_norm=3.517, loss_scale=1.000, learning_rate=1.444e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 07:40:26,222 (deepspeed_trainer:226) INFO: 8epoch:train:10101-10200batch: iter_time=1.122e-04, loss_ctc=84.436, loss_att=69.281, acc=0.634, loss=73.825, grad_norm=3.479, loss_scale=1.000, learning_rate=1.444e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 07:41:04,028 (deepspeed_trainer:226) INFO: 8epoch:train:10201-10300batch: iter_time=1.065e-04, loss_ctc=89.794, loss_att=70.485, acc=0.629, loss=76.254, grad_norm=4.285, loss_scale=1.000, learning_rate=1.443e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 07:41:41,817 (deepspeed_trainer:226) INFO: 8epoch:train:10301-10400batch: iter_time=1.094e-04, loss_ctc=97.987, loss_att=87.495, acc=0.602, loss=90.654, grad_norm=4.086, loss_scale=1.000, learning_rate=1.442e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 07:42:18,844 (deepspeed_trainer:226) INFO: 8epoch:train:10401-10500batch: iter_time=1.124e-04, loss_ctc=82.671, loss_att=68.681, acc=0.631, loss=72.881, grad_norm=3.358, loss_scale=1.000, learning_rate=1.442e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 07:42:56,632 (deepspeed_trainer:226) INFO: 8epoch:train:10501-10600batch: iter_time=1.108e-04, loss_ctc=90.581, loss_att=69.499, acc=0.629, loss=75.841, grad_norm=3.627, loss_scale=1.000, learning_rate=1.441e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 07:43:34,098 (deepspeed_trainer:226) INFO: 8epoch:train:10601-10700batch: iter_time=1.083e-04, loss_ctc=86.486, loss_att=69.438, acc=0.631, loss=74.561, grad_norm=3.468, loss_scale=1.000, learning_rate=1.441e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:44:11,584 (deepspeed_trainer:226) INFO: 8epoch:train:10701-10800batch: iter_time=1.105e-04, loss_ctc=95.029, loss_att=80.374, acc=0.634, loss=84.778, grad_norm=3.818, loss_scale=1.000, learning_rate=1.440e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:44:49,627 (deepspeed_trainer:226) INFO: 8epoch:train:10801-10900batch: iter_time=1.099e-04, loss_ctc=102.069, loss_att=73.979, acc=0.640, loss=82.393, grad_norm=4.136, loss_scale=1.000, learning_rate=1.439e-04, step_time=0.380 [2024-11-29 07:45:27,049] [INFO] [logging.py:129:log_dist] [Rank 0] step=116000, skipped=0, lr=[np.float64(0.00014383837045397533)], mom=[[0.9, 0.98]] [2024-11-29 07:45:27,050] [INFO] [timer.py:264:stop] epoch=0/micro_step=116000/global_step=116000, RunningAvgSamplesPerSec=43.02964579686061, CurrSamplesPerSec=45.50931506149935, MemAllocated=2.04GB, MaxMemAllocated=16.57GB [cnode7-008:0/16] 2024-11-29 07:45:27,051 (deepspeed_trainer:226) INFO: 8epoch:train:10901-11000batch: iter_time=1.127e-04, loss_ctc=82.659, loss_att=65.022, acc=0.645, loss=70.301, grad_norm=3.432, loss_scale=1.000, learning_rate=1.439e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:46:03,394 (deepspeed_trainer:226) INFO: 8epoch:train:11001-11100batch: iter_time=1.102e-04, loss_ctc=80.360, loss_att=59.957, acc=0.640, loss=66.064, grad_norm=3.343, loss_scale=1.000, learning_rate=1.438e-04, step_time=0.363 [cnode7-008:0/16] 2024-11-29 07:46:40,548 (deepspeed_trainer:226) INFO: 8epoch:train:11101-11200batch: iter_time=1.095e-04, loss_ctc=94.410, loss_att=82.047, acc=0.631, loss=85.781, grad_norm=3.930, loss_scale=1.000, learning_rate=1.437e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 07:47:02,932 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-29 07:47:30,325 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 07:47:36,008 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 07:47:36,008 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-29 07:47:36,011 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 07:47:51,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:03,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:53,000] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:53,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:53,329] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:03,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:53,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:04,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:04,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:53,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:54,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:47:54,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:04,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:05,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:05,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:05,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:40,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:40,730] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:41,245] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:41,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:52,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:42,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:42,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:53,150] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:42,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:42,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:53,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:53,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:54,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:54,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:54,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:48:54,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:27,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:28,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:29,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:29,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:29,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:30,213] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:30,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:41,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:30,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:42,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:42,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:42,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:44,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:44,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:45,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:49:45,070] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:14,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:15,790] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:16,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:17,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:17,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:18,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:29,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:18,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:30,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:20,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:31,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:32,113] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:32,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:32,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:32,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 07:50:33,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 07:51:20,908 (deepspeed_trainer:226) INFO: 8epoch:train:11201-11300batch: iter_time=2.406, loss_ctc=96.549, loss_att=82.036, acc=0.624, loss=86.379, grad_norm=3.870, loss_scale=1.000, learning_rate=1.437e-04, step_time=0.398 [cnode7-008:0/16] 2024-11-29 07:51:58,604 (deepspeed_trainer:226) INFO: 8epoch:train:11301-11400batch: iter_time=1.100e-04, loss_ctc=76.507, loss_att=57.025, acc=0.646, loss=62.863, grad_norm=3.170, loss_scale=1.000, learning_rate=1.436e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 07:52:35,678 (deepspeed_trainer:226) INFO: 8epoch:train:11401-11500batch: iter_time=1.116e-04, loss_ctc=78.748, loss_att=67.406, acc=0.613, loss=70.803, grad_norm=3.704, loss_scale=1.000, learning_rate=1.436e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 07:53:13,059 (deepspeed_trainer:226) INFO: 8epoch:train:11501-11600batch: iter_time=1.080e-04, loss_ctc=82.841, loss_att=74.277, acc=0.626, loss=76.823, grad_norm=3.299, loss_scale=1.000, learning_rate=1.435e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 07:53:51,249 (deepspeed_trainer:226) INFO: 8epoch:train:11601-11700batch: iter_time=1.075e-04, loss_ctc=93.520, loss_att=76.100, acc=0.625, loss=81.346, grad_norm=3.921, loss_scale=1.000, learning_rate=1.434e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 07:54:28,496 (deepspeed_trainer:226) INFO: 8epoch:train:11701-11800batch: iter_time=1.145e-04, loss_ctc=91.419, loss_att=74.114, acc=0.622, loss=79.304, grad_norm=3.725, loss_scale=1.000, learning_rate=1.434e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:55:06,796 (deepspeed_trainer:226) INFO: 8epoch:train:11801-11900batch: iter_time=1.101e-04, loss_ctc=90.980, loss_att=78.579, acc=0.612, loss=82.282, grad_norm=3.766, loss_scale=1.000, learning_rate=1.433e-04, step_time=0.378 [2024-11-29 07:55:44,637] [INFO] [logging.py:129:log_dist] [Rank 0] step=117000, skipped=0, lr=[np.float64(0.00014322236274781514)], mom=[[0.9, 0.98]] [2024-11-29 07:55:44,639] [INFO] [timer.py:264:stop] epoch=0/micro_step=117000/global_step=117000, RunningAvgSamplesPerSec=43.031869294920355, CurrSamplesPerSec=43.52127951477992, MemAllocated=2.04GB, MaxMemAllocated=16.57GB [cnode7-008:0/16] 2024-11-29 07:55:44,640 (deepspeed_trainer:226) INFO: 8epoch:train:11901-12000batch: iter_time=1.118e-04, loss_ctc=83.924, loss_att=65.757, acc=0.633, loss=71.204, grad_norm=3.264, loss_scale=1.000, learning_rate=1.433e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 07:56:21,869 (deepspeed_trainer:226) INFO: 8epoch:train:12001-12100batch: iter_time=1.109e-04, loss_ctc=89.045, loss_att=75.494, acc=0.624, loss=79.547, grad_norm=3.563, loss_scale=1.000, learning_rate=1.432e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 07:56:59,141 (deepspeed_trainer:226) INFO: 8epoch:train:12101-12200batch: iter_time=1.125e-04, loss_ctc=91.578, loss_att=73.758, acc=0.617, loss=79.098, grad_norm=3.938, loss_scale=1.000, learning_rate=1.431e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 07:57:36,118 (deepspeed_trainer:226) INFO: 8epoch:train:12201-12300batch: iter_time=1.096e-04, loss_ctc=90.142, loss_att=75.561, acc=0.618, loss=79.916, grad_norm=3.469, loss_scale=1.000, learning_rate=1.431e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 07:58:13,112 (deepspeed_trainer:226) INFO: 8epoch:train:12301-12400batch: iter_time=1.124e-04, loss_ctc=89.461, loss_att=72.018, acc=0.625, loss=77.215, grad_norm=3.914, loss_scale=1.000, learning_rate=1.430e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 07:58:50,583 (deepspeed_trainer:226) INFO: 8epoch:train:12401-12500batch: iter_time=1.123e-04, loss_ctc=91.440, loss_att=68.563, acc=0.629, loss=75.420, grad_norm=3.613, loss_scale=1.000, learning_rate=1.429e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 07:59:28,568 (deepspeed_trainer:226) INFO: 8epoch:train:12501-12600batch: iter_time=1.076e-04, loss_ctc=85.270, loss_att=71.056, acc=0.633, loss=75.306, grad_norm=3.276, loss_scale=1.000, learning_rate=1.429e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 08:00:05,928 (deepspeed_trainer:226) INFO: 8epoch:train:12601-12700batch: iter_time=1.091e-04, loss_ctc=95.784, loss_att=72.039, acc=0.633, loss=79.108, grad_norm=3.944, loss_scale=1.000, learning_rate=1.428e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 08:00:45,196 (deepspeed_trainer:226) INFO: 8epoch:train:12701-12800batch: iter_time=1.093e-04, loss_ctc=101.451, loss_att=72.690, acc=0.645, loss=81.332, grad_norm=4.355, loss_scale=1.000, learning_rate=1.428e-04, step_time=0.392 [cnode7-008:0/16] 2024-11-29 08:01:22,325 (deepspeed_trainer:226) INFO: 8epoch:train:12801-12900batch: iter_time=1.092e-04, loss_ctc=72.474, loss_att=56.993, acc=0.644, loss=61.632, grad_norm=3.187, loss_scale=1.000, learning_rate=1.427e-04, step_time=0.371 [2024-11-29 08:01:59,379] [INFO] [logging.py:129:log_dist] [Rank 0] step=118000, skipped=0, lr=[np.float64(0.00014261420226627158)], mom=[[0.9, 0.98]] [2024-11-29 08:01:59,379] [INFO] [timer.py:264:stop] epoch=0/micro_step=118000/global_step=118000, RunningAvgSamplesPerSec=43.03632596010568, CurrSamplesPerSec=40.41674715636408, MemAllocated=2.04GB, MaxMemAllocated=16.57GB [cnode7-008:0/16] 2024-11-29 08:01:59,381 (deepspeed_trainer:226) INFO: 8epoch:train:12901-13000batch: iter_time=1.077e-04, loss_ctc=90.951, loss_att=70.279, acc=0.632, loss=76.473, grad_norm=4.222, loss_scale=1.000, learning_rate=1.426e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:02:38,014 (deepspeed_trainer:226) INFO: 8epoch:train:13001-13100batch: iter_time=1.078e-04, loss_ctc=89.221, loss_att=74.486, acc=0.634, loss=78.917, grad_norm=3.609, loss_scale=1.000, learning_rate=1.426e-04, step_time=0.386 [cnode7-008:0/16] 2024-11-29 08:02:51,107 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-008:0/16] 2024-11-29 08:03:18,128 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 08:03:23,667 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 08:03:23,667 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-29 08:03:23,670 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 08:03:39,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:39,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:40,078] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:51,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:40,519] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:40,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:41,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:51,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:41,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:41,361] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:52,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:52,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:53,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:53,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:53,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:03:53,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:28,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:28,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:40,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:29,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:29,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:29,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:29,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:29,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:41,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:41,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:41,774] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:42,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:43,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:43,655] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:44,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:04:33,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:16,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:16,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:18,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:29,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:18,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:18,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:18,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:30,322] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:30,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:30,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:31,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:32,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:21,873] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:33,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:33,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:05:24,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:03,857] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:04,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:05,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:07,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:07,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:07,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:18,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:19,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:20,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:21,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:21,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:21,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:22,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:22,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:12,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:06:14,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 08:07:19,112 (deepspeed_trainer:226) INFO: 8epoch:train:13101-13200batch: iter_time=2.402, loss_ctc=89.141, loss_att=77.581, acc=0.620, loss=81.051, grad_norm=3.525, loss_scale=1.000, learning_rate=1.425e-04, step_time=0.409 [cnode7-008:0/16] 2024-11-29 08:07:56,206 (deepspeed_trainer:226) INFO: 8epoch:train:13201-13300batch: iter_time=1.050e-04, loss_ctc=77.956, loss_att=55.667, acc=0.653, loss=62.383, grad_norm=3.350, loss_scale=1.000, learning_rate=1.425e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 08:08:33,472 (deepspeed_trainer:226) INFO: 8epoch:train:13301-13400batch: iter_time=1.156e-04, loss_ctc=82.902, loss_att=73.834, acc=0.622, loss=76.562, grad_norm=3.519, loss_scale=1.000, learning_rate=1.424e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 08:09:10,992 (deepspeed_trainer:226) INFO: 8epoch:train:13401-13500batch: iter_time=1.104e-04, loss_ctc=84.879, loss_att=74.272, acc=0.644, loss=77.459, grad_norm=3.460, loss_scale=1.000, learning_rate=1.423e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 08:09:48,364 (deepspeed_trainer:226) INFO: 8epoch:train:13501-13600batch: iter_time=1.111e-04, loss_ctc=88.471, loss_att=69.183, acc=0.635, loss=74.979, grad_norm=4.029, loss_scale=1.000, learning_rate=1.423e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 08:10:25,676 (deepspeed_trainer:226) INFO: 8epoch:train:13601-13700batch: iter_time=1.085e-04, loss_ctc=90.534, loss_att=76.689, acc=0.629, loss=80.851, grad_norm=3.924, loss_scale=1.000, learning_rate=1.422e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 08:11:04,010 (deepspeed_trainer:226) INFO: 8epoch:train:13701-13800batch: iter_time=1.093e-04, loss_ctc=90.945, loss_att=78.182, acc=0.614, loss=82.007, grad_norm=3.825, loss_scale=1.000, learning_rate=1.422e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 08:11:42,141 (deepspeed_trainer:226) INFO: 8epoch:train:13801-13900batch: iter_time=1.131e-04, loss_ctc=87.897, loss_att=70.154, acc=0.645, loss=75.474, grad_norm=3.640, loss_scale=1.000, learning_rate=1.421e-04, step_time=0.381 [2024-11-29 08:12:19,638] [INFO] [logging.py:129:log_dist] [Rank 0] step=119000, skipped=0, lr=[np.float64(0.00014201372380428032)], mom=[[0.9, 0.98]] [2024-11-29 08:12:19,638] [INFO] [timer.py:264:stop] epoch=0/micro_step=119000/global_step=119000, RunningAvgSamplesPerSec=43.035520551717624, CurrSamplesPerSec=46.79519928394356, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:12:19,640 (deepspeed_trainer:226) INFO: 8epoch:train:13901-14000batch: iter_time=1.104e-04, loss_ctc=85.084, loss_att=71.169, acc=0.627, loss=75.340, grad_norm=3.438, loss_scale=1.000, learning_rate=1.420e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 08:12:56,725 (deepspeed_trainer:226) INFO: 8epoch:train:14001-14100batch: iter_time=1.136e-04, loss_ctc=94.255, loss_att=78.123, acc=0.611, loss=82.986, grad_norm=4.095, loss_scale=1.000, learning_rate=1.420e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:13:34,929 (deepspeed_trainer:226) INFO: 8epoch:train:14101-14200batch: iter_time=1.113e-04, loss_ctc=84.517, loss_att=70.178, acc=0.636, loss=74.498, grad_norm=3.492, loss_scale=1.000, learning_rate=1.419e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 08:14:12,141 (deepspeed_trainer:226) INFO: 8epoch:train:14201-14300batch: iter_time=1.096e-04, loss_ctc=91.983, loss_att=77.327, acc=0.626, loss=81.711, grad_norm=3.888, loss_scale=1.000, learning_rate=1.419e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 08:14:49,502 (deepspeed_trainer:226) INFO: 8epoch:train:14301-14400batch: iter_time=1.172e-04, loss_ctc=87.551, loss_att=68.487, acc=0.628, loss=74.194, grad_norm=3.556, loss_scale=1.000, learning_rate=1.418e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 08:15:27,147 (deepspeed_trainer:226) INFO: 8epoch:train:14401-14500batch: iter_time=1.141e-04, loss_ctc=86.537, loss_att=70.877, acc=0.640, loss=75.585, grad_norm=3.589, loss_scale=1.000, learning_rate=1.417e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:16:04,636 (deepspeed_trainer:226) INFO: 8epoch:train:14501-14600batch: iter_time=1.113e-04, loss_ctc=97.340, loss_att=74.267, acc=0.643, loss=81.202, grad_norm=3.935, loss_scale=1.000, learning_rate=1.417e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:16:42,255 (deepspeed_trainer:226) INFO: 8epoch:train:14601-14700batch: iter_time=1.098e-04, loss_ctc=97.611, loss_att=73.137, acc=0.652, loss=80.518, grad_norm=4.390, loss_scale=1.000, learning_rate=1.416e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:17:19,240 (deepspeed_trainer:226) INFO: 8epoch:train:14701-14800batch: iter_time=1.136e-04, loss_ctc=72.731, loss_att=56.851, acc=0.647, loss=61.621, grad_norm=3.160, loss_scale=1.000, learning_rate=1.416e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:17:56,433 (deepspeed_trainer:226) INFO: 8epoch:train:14801-14900batch: iter_time=1.097e-04, loss_ctc=91.320, loss_att=71.293, acc=0.639, loss=77.309, grad_norm=3.815, loss_scale=1.000, learning_rate=1.415e-04, step_time=0.372 [2024-11-29 08:18:34,645] [INFO] [logging.py:129:log_dist] [Rank 0] step=120000, skipped=0, lr=[np.float64(0.00014142076698534133)], mom=[[0.9, 0.98]] [2024-11-29 08:18:34,646] [INFO] [timer.py:264:stop] epoch=0/micro_step=120000/global_step=120000, RunningAvgSamplesPerSec=43.03968262306855, CurrSamplesPerSec=42.48640946348247, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:18:34,647 (deepspeed_trainer:226) INFO: 8epoch:train:14901-15000batch: iter_time=1.093e-04, loss_ctc=90.539, loss_att=76.258, acc=0.636, loss=80.531, grad_norm=3.752, loss_scale=1.000, learning_rate=1.414e-04, step_time=0.382 [2024-11-29 08:18:50,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:40,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:40,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:40,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:41,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:41,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:41,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:41,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:41,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:50,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:56,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:56,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:06,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:56,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:07,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:07,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:07,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:58,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:58,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:07,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:58,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:58,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:18:58,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:08,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:08,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:08,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:11,994] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:12,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:12,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:22,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:12,929] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:13,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:23,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:13,609] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:14,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:14,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,916] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:24,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:27,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:37,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:28,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:28,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:29,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:38,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:29,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:29,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:30,042] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:30,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:39,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:40,154] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:40,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:40,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:41,033] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:41,067] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:19:53,694] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 8 is about to be saved! [2024-11-29 08:19:53,724] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/mp_rank_00_model_states.pt [2024-11-29 08:19:53,724] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/mp_rank_00_model_states.pt... [2024-11-29 08:19:55,629] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/mp_rank_00_model_states.pt. [2024-11-29 08:19:55,785] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,245] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,785] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,245] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,247] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,247] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,788] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,791] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,252] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,792] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,792] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,793] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,253] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-11-29 08:19:55,793] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,253] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-11-29 08:19:46,253] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-11-29 08:19:56,363] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,364] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,364] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,478] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,481] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,482] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:46,956] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-11-29 08:19:46,956] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-11-29 08:19:46,956] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,005] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,006] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,006] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,567] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,567] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,567] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,575] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,575] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,575] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,037] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,037] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,037] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,037] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,037] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,037] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,581] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,582] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,582] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,043] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,043] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,043] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,588] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,588] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,588] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,598] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,599] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,599] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:56,600] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-11-29 08:19:56,600] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-11-29 08:19:56,600] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,069] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,069] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,069] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,069] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,069] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,069] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [2024-11-29 08:19:47,074] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-11-29 08:19:47,074] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_8/8/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-11-29 08:19:47,074] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 8 is ready now! [cnode7-008:0/16] 2024-11-29 08:19:56,625 (deepspeed_trainer:158) INFO: 8epoch results: [train] iter_time=0.129, loss_ctc=90.319, loss_att=73.386, acc=0.623, loss=78.467, grad_norm=3.728, loss_scale=1.000, learning_rate=1.461e-04, step_time=0.378, time=2 hours, 6 minutes and 51.64 seconds, total_count=120008, gpu_max_cached_mem_GB=33.807, [valid] loss_ctc=8.312, cer_ctc=0.257, loss_att=10.000, acc=0.847, cer=0.113, wer=1.000, loss=9.500, time=1 minute and 10.16 seconds, total_count=8, gpu_max_cached_mem_GB=33.807 [cnode7-008:0/16] 2024-11-29 08:19:58,823 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-008:0/16] 2024-11-29 08:20:25,490 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 08:20:30,944 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.4", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.4", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.4", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.4", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 08:20:30,944 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, [cnode7-008:0/16] 2024-11-29 08:20:30,946 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 08:20:42,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:52,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:53,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:43,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:53,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:44,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:54,275] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:54,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:54,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:54,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:20:55,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:30,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:31,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:31,170] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:31,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:31,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:32,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:32,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:32,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:42,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:43,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:43,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:44,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:44,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:45,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:45,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:21:45,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:17,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:17,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:18,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:19,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:19,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:19,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:29,485] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:20,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:20,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:31,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:31,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:31,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:32,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:33,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:33,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:22:35,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:05,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:05,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:05,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:06,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:07,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:07,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:07,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:17,883] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:21,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:11,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:21,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:21,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:22,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:22,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:23,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:23:26,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 08:24:30,259 (deepspeed_trainer:226) INFO: 9epoch:train:1-100batch: iter_time=2.340, loss_ctc=89.064, loss_att=65.419, acc=0.631, loss=72.515, grad_norm=3.853, loss_scale=1.000, learning_rate=1.414e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 08:25:08,642 (deepspeed_trainer:226) INFO: 9epoch:train:101-200batch: iter_time=1.052e-04, loss_ctc=85.167, loss_att=66.981, acc=0.625, loss=72.452, grad_norm=3.920, loss_scale=1.000, learning_rate=1.413e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 08:25:45,901 (deepspeed_trainer:226) INFO: 9epoch:train:201-300batch: iter_time=1.094e-04, loss_ctc=89.937, loss_att=66.892, acc=0.632, loss=73.776, grad_norm=3.877, loss_scale=1.000, learning_rate=1.413e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 08:26:23,489 (deepspeed_trainer:226) INFO: 9epoch:train:301-400batch: iter_time=1.071e-04, loss_ctc=84.862, loss_att=70.639, acc=0.636, loss=74.895, grad_norm=3.762, loss_scale=1.000, learning_rate=1.412e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:27:01,387 (deepspeed_trainer:226) INFO: 9epoch:train:401-500batch: iter_time=1.077e-04, loss_ctc=85.709, loss_att=77.250, acc=0.602, loss=79.812, grad_norm=3.676, loss_scale=1.000, learning_rate=1.412e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 08:27:39,256 (deepspeed_trainer:226) INFO: 9epoch:train:501-600batch: iter_time=1.065e-04, loss_ctc=90.284, loss_att=68.258, acc=0.619, loss=74.873, grad_norm=3.960, loss_scale=1.000, learning_rate=1.411e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 08:28:16,903 (deepspeed_trainer:226) INFO: 9epoch:train:601-700batch: iter_time=1.059e-04, loss_ctc=93.318, loss_att=72.256, acc=0.645, loss=78.579, grad_norm=4.272, loss_scale=1.000, learning_rate=1.410e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:28:54,079 (deepspeed_trainer:226) INFO: 9epoch:train:701-800batch: iter_time=1.059e-04, loss_ctc=106.072, loss_att=75.190, acc=0.636, loss=84.436, grad_norm=4.675, loss_scale=1.000, learning_rate=1.410e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 08:29:31,525 (deepspeed_trainer:226) INFO: 9epoch:train:801-900batch: iter_time=1.046e-04, loss_ctc=94.093, loss_att=72.183, acc=0.631, loss=78.730, grad_norm=3.778, loss_scale=1.000, learning_rate=1.409e-04, step_time=0.374 [2024-11-29 08:30:09,730] [INFO] [logging.py:129:log_dist] [Rank 0] step=121000, skipped=0, lr=[np.float64(0.00014083517608157094)], mom=[[0.9, 0.98]] [2024-11-29 08:30:09,731] [INFO] [timer.py:264:stop] epoch=0/micro_step=121000/global_step=121000, RunningAvgSamplesPerSec=43.04151564263667, CurrSamplesPerSec=44.30501821241139, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:30:09,732 (deepspeed_trainer:226) INFO: 9epoch:train:901-1000batch: iter_time=1.039e-04, loss_ctc=84.706, loss_att=66.868, acc=0.648, loss=72.259, grad_norm=3.579, loss_scale=1.000, learning_rate=1.409e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 08:30:47,207 (deepspeed_trainer:226) INFO: 9epoch:train:1001-1100batch: iter_time=1.081e-04, loss_ctc=85.506, loss_att=71.962, acc=0.626, loss=76.021, grad_norm=3.887, loss_scale=1.000, learning_rate=1.408e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:31:25,558 (deepspeed_trainer:226) INFO: 9epoch:train:1101-1200batch: iter_time=1.044e-04, loss_ctc=84.034, loss_att=69.856, acc=0.638, loss=74.123, grad_norm=3.665, loss_scale=1.000, learning_rate=1.407e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 08:32:03,240 (deepspeed_trainer:226) INFO: 9epoch:train:1201-1300batch: iter_time=1.043e-04, loss_ctc=91.672, loss_att=73.899, acc=0.617, loss=79.217, grad_norm=3.781, loss_scale=1.000, learning_rate=1.407e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 08:32:40,076 (deepspeed_trainer:226) INFO: 9epoch:train:1301-1400batch: iter_time=1.035e-04, loss_ctc=95.284, loss_att=76.403, acc=0.625, loss=82.040, grad_norm=3.946, loss_scale=1.000, learning_rate=1.406e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 08:33:18,383 (deepspeed_trainer:226) INFO: 9epoch:train:1401-1500batch: iter_time=1.042e-04, loss_ctc=92.875, loss_att=72.737, acc=0.636, loss=78.816, grad_norm=4.237, loss_scale=1.000, learning_rate=1.406e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 08:33:55,367 (deepspeed_trainer:226) INFO: 9epoch:train:1501-1600batch: iter_time=1.045e-04, loss_ctc=97.947, loss_att=80.332, acc=0.633, loss=85.583, grad_norm=3.966, loss_scale=1.000, learning_rate=1.405e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:34:33,021 (deepspeed_trainer:226) INFO: 9epoch:train:1601-1700batch: iter_time=1.042e-04, loss_ctc=88.063, loss_att=69.384, acc=0.649, loss=75.012, grad_norm=3.793, loss_scale=1.000, learning_rate=1.405e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:35:10,121 (deepspeed_trainer:226) INFO: 9epoch:train:1701-1800batch: iter_time=1.069e-04, loss_ctc=89.632, loss_att=73.874, acc=0.626, loss=78.582, grad_norm=3.833, loss_scale=1.000, learning_rate=1.404e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 08:35:43,136 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-008:0/16] 2024-11-29 08:36:10,685 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 08:36:16,587 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.5", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.5", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.5", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.5", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 08:36:16,587 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, [cnode7-008:0/16] 2024-11-29 08:36:16,589 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 08:36:36,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:36,567] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:36,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:36,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:37,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:37,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:37,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:37,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:36:52,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:27,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:27,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:28,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:29,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:29,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:29,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:29,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:29,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:48,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:48,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:49,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:49,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:49,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:50,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:50,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:37:50,157] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:15,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:15,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:16,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:17,236] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:17,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:19,243] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:20,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:21,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:36,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:38,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:38:39,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:03,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:03,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:03,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:04,256] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:05,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:08,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:10,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:11,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:25,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:27,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:27,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:28,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:28,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:28,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:29,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:39:29,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 08:40:06,416 (deepspeed_trainer:226) INFO: 9epoch:train:1801-1900batch: iter_time=2.587, loss_ctc=85.633, loss_att=68.486, acc=0.643, loss=73.606, grad_norm=3.590, loss_scale=1.000, learning_rate=1.403e-04, step_time=0.375 [2024-11-29 08:40:43,814] [INFO] [logging.py:129:log_dist] [Rank 0] step=122000, skipped=0, lr=[np.float64(0.00014025679984188308)], mom=[[0.9, 0.98]] [2024-11-29 08:40:43,817] [INFO] [timer.py:264:stop] epoch=0/micro_step=122000/global_step=122000, RunningAvgSamplesPerSec=43.044799856392316, CurrSamplesPerSec=45.587901128806166, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:40:43,818 (deepspeed_trainer:226) INFO: 9epoch:train:1901-2000batch: iter_time=1.047e-04, loss_ctc=86.660, loss_att=64.564, acc=0.633, loss=71.192, grad_norm=3.840, loss_scale=1.000, learning_rate=1.403e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:41:22,274 (deepspeed_trainer:226) INFO: 9epoch:train:2001-2100batch: iter_time=1.087e-04, loss_ctc=83.800, loss_att=66.591, acc=0.628, loss=71.799, grad_norm=3.698, loss_scale=1.000, learning_rate=1.402e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 08:41:59,841 (deepspeed_trainer:226) INFO: 9epoch:train:2101-2200batch: iter_time=1.079e-04, loss_ctc=88.385, loss_att=67.899, acc=0.621, loss=74.051, grad_norm=4.051, loss_scale=1.000, learning_rate=1.402e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 08:42:38,264 (deepspeed_trainer:226) INFO: 9epoch:train:2201-2300batch: iter_time=1.063e-04, loss_ctc=86.239, loss_att=72.475, acc=0.636, loss=76.634, grad_norm=3.580, loss_scale=1.000, learning_rate=1.401e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 08:43:15,240 (deepspeed_trainer:226) INFO: 9epoch:train:2301-2400batch: iter_time=1.094e-04, loss_ctc=82.894, loss_att=71.513, acc=0.609, loss=74.932, grad_norm=3.493, loss_scale=1.000, learning_rate=1.401e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:43:52,374 (deepspeed_trainer:226) INFO: 9epoch:train:2401-2500batch: iter_time=1.083e-04, loss_ctc=89.742, loss_att=67.255, acc=0.626, loss=73.998, grad_norm=4.137, loss_scale=1.000, learning_rate=1.400e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 08:44:30,346 (deepspeed_trainer:226) INFO: 9epoch:train:2501-2600batch: iter_time=1.090e-04, loss_ctc=92.344, loss_att=71.407, acc=0.634, loss=77.643, grad_norm=3.827, loss_scale=1.000, learning_rate=1.399e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 08:45:08,541 (deepspeed_trainer:226) INFO: 9epoch:train:2601-2700batch: iter_time=1.149e-04, loss_ctc=110.823, loss_att=74.949, acc=0.643, loss=85.728, grad_norm=5.032, loss_scale=1.000, learning_rate=1.399e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 08:45:45,660 (deepspeed_trainer:226) INFO: 9epoch:train:2701-2800batch: iter_time=1.097e-04, loss_ctc=87.006, loss_att=64.448, acc=0.632, loss=71.196, grad_norm=3.475, loss_scale=1.000, learning_rate=1.398e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 08:46:23,075 (deepspeed_trainer:226) INFO: 9epoch:train:2801-2900batch: iter_time=1.119e-04, loss_ctc=80.875, loss_att=68.435, acc=0.641, loss=72.192, grad_norm=3.346, loss_scale=1.000, learning_rate=1.398e-04, step_time=0.374 [2024-11-29 08:47:00,061] [INFO] [logging.py:129:log_dist] [Rank 0] step=123000, skipped=0, lr=[np.float64(0.00013968549132786922)], mom=[[0.9, 0.98]] [2024-11-29 08:47:00,061] [INFO] [timer.py:264:stop] epoch=0/micro_step=123000/global_step=123000, RunningAvgSamplesPerSec=43.04743991454924, CurrSamplesPerSec=42.34208252194526, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:47:00,063 (deepspeed_trainer:226) INFO: 9epoch:train:2901-3000batch: iter_time=1.105e-04, loss_ctc=85.396, loss_att=72.797, acc=0.628, loss=76.528, grad_norm=3.488, loss_scale=1.000, learning_rate=1.397e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 08:47:37,443 (deepspeed_trainer:226) INFO: 9epoch:train:3001-3100batch: iter_time=1.119e-04, loss_ctc=89.293, loss_att=70.407, acc=0.626, loss=76.045, grad_norm=3.572, loss_scale=1.000, learning_rate=1.397e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:48:14,937 (deepspeed_trainer:226) INFO: 9epoch:train:3101-3200batch: iter_time=1.126e-04, loss_ctc=90.416, loss_att=71.912, acc=0.629, loss=77.470, grad_norm=3.725, loss_scale=1.000, learning_rate=1.396e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:48:53,179 (deepspeed_trainer:226) INFO: 9epoch:train:3201-3300batch: iter_time=1.109e-04, loss_ctc=88.743, loss_att=75.224, acc=0.621, loss=79.296, grad_norm=3.990, loss_scale=1.000, learning_rate=1.395e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 08:49:31,202 (deepspeed_trainer:226) INFO: 9epoch:train:3301-3400batch: iter_time=1.111e-04, loss_ctc=95.736, loss_att=72.945, acc=0.640, loss=79.793, grad_norm=4.052, loss_scale=1.000, learning_rate=1.395e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 08:50:08,761 (deepspeed_trainer:226) INFO: 9epoch:train:3401-3500batch: iter_time=1.094e-04, loss_ctc=91.153, loss_att=72.126, acc=0.638, loss=77.841, grad_norm=3.856, loss_scale=1.000, learning_rate=1.394e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 08:50:46,968 (deepspeed_trainer:226) INFO: 9epoch:train:3501-3600batch: iter_time=1.069e-04, loss_ctc=89.964, loss_att=69.489, acc=0.640, loss=75.667, grad_norm=3.904, loss_scale=1.000, learning_rate=1.394e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 08:51:24,000 (deepspeed_trainer:226) INFO: 9epoch:train:3601-3700batch: iter_time=1.114e-04, loss_ctc=88.954, loss_att=74.539, acc=0.627, loss=78.881, grad_norm=3.803, loss_scale=1.000, learning_rate=1.393e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 08:51:47,087 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-008:0/16] 2024-11-29 08:52:14,353 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 08:52:20,319 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.6", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.6", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.6", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.6", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 08:52:20,319 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, [cnode7-008:0/16] 2024-11-29 08:52:20,322 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 08:52:36,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:36,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:36,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:36,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:37,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:37,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:37,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:37,970] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:48,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:48,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:48,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:49,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:49,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:50,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:50,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:52:50,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:24,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:24,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:25,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:25,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:26,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:26,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:26,687] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:37,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:37,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:38,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:38,801] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:39,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:30,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:40,542] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:40,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:53:40,615] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:12,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:13,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:13,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:14,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:14,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:14,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:26,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:27,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:27,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:18,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:29,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:29,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:30,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:30,737] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:20,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:54:34,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:00,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:01,184] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:01,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:02,170] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:02,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:02,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:14,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:16,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:16,448] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:17,811] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:18,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:08,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:18,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:19,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:10,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 08:55:24,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 08:56:09,042 (deepspeed_trainer:226) INFO: 9epoch:train:3701-3800batch: iter_time=2.476, loss_ctc=87.216, loss_att=68.843, acc=0.644, loss=74.350, grad_norm=3.769, loss_scale=1.000, learning_rate=1.393e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 08:56:46,121 (deepspeed_trainer:226) INFO: 9epoch:train:3801-3900batch: iter_time=1.094e-04, loss_ctc=78.457, loss_att=58.495, acc=0.648, loss=64.474, grad_norm=3.491, loss_scale=1.000, learning_rate=1.392e-04, step_time=0.370 [2024-11-29 08:57:23,097] [INFO] [logging.py:129:log_dist] [Rank 0] step=124000, skipped=0, lr=[np.float64(0.00013912110775697278)], mom=[[0.9, 0.98]] [2024-11-29 08:57:23,097] [INFO] [timer.py:264:stop] epoch=0/micro_step=124000/global_step=124000, RunningAvgSamplesPerSec=43.0507463614999, CurrSamplesPerSec=42.95293891443391, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 08:57:23,098 (deepspeed_trainer:226) INFO: 9epoch:train:3901-4000batch: iter_time=1.087e-04, loss_ctc=85.063, loss_att=68.186, acc=0.638, loss=73.252, grad_norm=3.592, loss_scale=1.000, learning_rate=1.391e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 08:58:00,676 (deepspeed_trainer:226) INFO: 9epoch:train:4001-4100batch: iter_time=1.123e-04, loss_ctc=87.053, loss_att=71.931, acc=0.620, loss=76.458, grad_norm=3.823, loss_scale=1.000, learning_rate=1.391e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 08:58:38,561 (deepspeed_trainer:226) INFO: 9epoch:train:4101-4200batch: iter_time=1.095e-04, loss_ctc=83.716, loss_att=65.467, acc=0.637, loss=70.936, grad_norm=3.607, loss_scale=1.000, learning_rate=1.390e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 08:59:16,946 (deepspeed_trainer:226) INFO: 9epoch:train:4201-4300batch: iter_time=1.106e-04, loss_ctc=87.999, loss_att=75.568, acc=0.617, loss=79.337, grad_norm=3.918, loss_scale=1.000, learning_rate=1.390e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 08:59:54,111 (deepspeed_trainer:226) INFO: 9epoch:train:4301-4400batch: iter_time=1.092e-04, loss_ctc=88.158, loss_att=67.436, acc=0.641, loss=73.624, grad_norm=3.950, loss_scale=1.000, learning_rate=1.389e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 09:00:32,927 (deepspeed_trainer:226) INFO: 9epoch:train:4401-4500batch: iter_time=1.066e-04, loss_ctc=97.612, loss_att=72.103, acc=0.644, loss=79.817, grad_norm=4.062, loss_scale=1.000, learning_rate=1.389e-04, step_time=0.388 [cnode7-008:0/16] 2024-11-29 09:01:10,011 (deepspeed_trainer:226) INFO: 9epoch:train:4501-4600batch: iter_time=1.087e-04, loss_ctc=95.582, loss_att=70.369, acc=0.654, loss=77.934, grad_norm=4.801, loss_scale=1.000, learning_rate=1.388e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 09:01:47,622 (deepspeed_trainer:226) INFO: 9epoch:train:4601-4700batch: iter_time=1.086e-04, loss_ctc=86.751, loss_att=65.367, acc=0.637, loss=71.784, grad_norm=3.694, loss_scale=1.000, learning_rate=1.388e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 09:02:25,551 (deepspeed_trainer:226) INFO: 9epoch:train:4701-4800batch: iter_time=1.084e-04, loss_ctc=82.647, loss_att=72.097, acc=0.647, loss=75.299, grad_norm=3.422, loss_scale=1.000, learning_rate=1.387e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 09:03:03,241 (deepspeed_trainer:226) INFO: 9epoch:train:4801-4900batch: iter_time=1.112e-04, loss_ctc=84.300, loss_att=69.783, acc=0.638, loss=74.147, grad_norm=3.371, loss_scale=1.000, learning_rate=1.386e-04, step_time=0.376 [2024-11-29 09:03:41,291] [INFO] [logging.py:129:log_dist] [Rank 0] step=125000, skipped=0, lr=[np.float64(0.00013856351035257728)], mom=[[0.9, 0.98]] [2024-11-29 09:03:41,291] [INFO] [timer.py:264:stop] epoch=0/micro_step=125000/global_step=125000, RunningAvgSamplesPerSec=43.05139519548506, CurrSamplesPerSec=44.10289103990349, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:03:41,293 (deepspeed_trainer:226) INFO: 9epoch:train:4901-5000batch: iter_time=1.136e-04, loss_ctc=90.091, loss_att=69.487, acc=0.634, loss=75.644, grad_norm=3.695, loss_scale=1.000, learning_rate=1.386e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 09:04:18,958 (deepspeed_trainer:226) INFO: 9epoch:train:5001-5100batch: iter_time=1.115e-04, loss_ctc=92.662, loss_att=73.150, acc=0.632, loss=79.034, grad_norm=3.920, loss_scale=1.000, learning_rate=1.385e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 09:04:56,023 (deepspeed_trainer:226) INFO: 9epoch:train:5101-5200batch: iter_time=1.095e-04, loss_ctc=86.061, loss_att=71.756, acc=0.634, loss=76.062, grad_norm=3.604, loss_scale=1.000, learning_rate=1.385e-04, step_time=0.370 [cnode7-008:0/16] 2024-11-29 09:05:33,135 (deepspeed_trainer:226) INFO: 9epoch:train:5201-5300batch: iter_time=1.096e-04, loss_ctc=88.706, loss_att=67.588, acc=0.654, loss=73.911, grad_norm=3.627, loss_scale=1.000, learning_rate=1.384e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 09:06:11,157 (deepspeed_trainer:226) INFO: 9epoch:train:5301-5400batch: iter_time=1.083e-04, loss_ctc=92.687, loss_att=77.733, acc=0.638, loss=82.231, grad_norm=3.992, loss_scale=1.000, learning_rate=1.384e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 09:06:48,901 (deepspeed_trainer:226) INFO: 9epoch:train:5401-5500batch: iter_time=1.096e-04, loss_ctc=88.033, loss_att=68.083, acc=0.643, loss=74.066, grad_norm=3.603, loss_scale=1.000, learning_rate=1.383e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 09:07:26,649 (deepspeed_trainer:226) INFO: 9epoch:train:5501-5600batch: iter_time=1.084e-04, loss_ctc=87.566, loss_att=74.075, acc=0.639, loss=78.091, grad_norm=3.788, loss_scale=1.000, learning_rate=1.383e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 09:07:40,298 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-008:0/16] 2024-11-29 09:08:07,199 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 09:08:13,106 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.2", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.2", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.2", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.2", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 09:08:13,106 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, [cnode7-008:0/16] 2024-11-29 09:08:13,108 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 09:08:28,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:28,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:29,321] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:39,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:29,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:40,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:29,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:30,144] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:30,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:30,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:40,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:40,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:41,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:41,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:41,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:08:42,143] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:16,698] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:17,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:17,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:18,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:18,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:29,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:19,046] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:19,085] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:29,578] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:30,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:30,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:31,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:21,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:31,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:31,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:09:32,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:04,446] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:05,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:06,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:06,714] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:07,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:07,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:18,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:19,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:20,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:20,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:20,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:20,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:20,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:10,296] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:10,783] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:25,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:51,965] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:52,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:54,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:55,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:55,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:55,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:08,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:08,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:09,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:09,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:09,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:10,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:10:59,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:10,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:00,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:11:16,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 09:12:12,067 (deepspeed_trainer:226) INFO: 9epoch:train:5601-5700batch: iter_time=2.416, loss_ctc=90.111, loss_att=67.022, acc=0.645, loss=73.980, grad_norm=3.728, loss_scale=1.000, learning_rate=1.382e-04, step_time=0.438 [cnode7-008:0/16] 2024-11-29 09:12:50,267 (deepspeed_trainer:226) INFO: 9epoch:train:5701-5800batch: iter_time=1.054e-04, loss_ctc=75.787, loss_att=58.359, acc=0.648, loss=63.595, grad_norm=3.642, loss_scale=1.000, learning_rate=1.381e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 09:13:28,550 (deepspeed_trainer:226) INFO: 9epoch:train:5801-5900batch: iter_time=1.071e-04, loss_ctc=91.980, loss_att=69.694, acc=0.642, loss=76.376, grad_norm=3.821, loss_scale=1.000, learning_rate=1.381e-04, step_time=0.383 [2024-11-29 09:14:05,834] [INFO] [logging.py:129:log_dist] [Rank 0] step=126000, skipped=0, lr=[np.float64(0.00013801256420065193)], mom=[[0.9, 0.98]] [2024-11-29 09:14:05,835] [INFO] [timer.py:264:stop] epoch=0/micro_step=126000/global_step=126000, RunningAvgSamplesPerSec=43.04777194513878, CurrSamplesPerSec=41.229380190371366, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:14:05,836 (deepspeed_trainer:226) INFO: 9epoch:train:5901-6000batch: iter_time=1.047e-04, loss_ctc=78.615, loss_att=66.710, acc=0.635, loss=70.270, grad_norm=3.541, loss_scale=1.000, learning_rate=1.380e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:14:43,236 (deepspeed_trainer:226) INFO: 9epoch:train:6001-6100batch: iter_time=1.080e-04, loss_ctc=81.220, loss_att=65.134, acc=0.631, loss=69.964, grad_norm=3.558, loss_scale=1.000, learning_rate=1.380e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:15:20,689 (deepspeed_trainer:226) INFO: 9epoch:train:6101-6200batch: iter_time=1.059e-04, loss_ctc=91.832, loss_att=75.369, acc=0.624, loss=80.308, grad_norm=3.953, loss_scale=1.000, learning_rate=1.379e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:15:58,295 (deepspeed_trainer:226) INFO: 9epoch:train:6201-6300batch: iter_time=1.095e-04, loss_ctc=88.605, loss_att=67.522, acc=0.648, loss=73.865, grad_norm=3.902, loss_scale=1.000, learning_rate=1.379e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 09:16:36,552 (deepspeed_trainer:226) INFO: 9epoch:train:6301-6400batch: iter_time=1.085e-04, loss_ctc=98.814, loss_att=72.108, acc=0.647, loss=80.129, grad_norm=3.852, loss_scale=1.000, learning_rate=1.378e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 09:17:14,955 (deepspeed_trainer:226) INFO: 9epoch:train:6401-6500batch: iter_time=1.066e-04, loss_ctc=90.587, loss_att=66.530, acc=0.652, loss=73.714, grad_norm=4.112, loss_scale=1.000, learning_rate=1.378e-04, step_time=0.384 [cnode7-008:0/16] 2024-11-29 09:17:52,521 (deepspeed_trainer:226) INFO: 9epoch:train:6501-6600batch: iter_time=1.163e-04, loss_ctc=84.099, loss_att=64.972, acc=0.652, loss=70.736, grad_norm=3.649, loss_scale=1.000, learning_rate=1.377e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:18:29,679 (deepspeed_trainer:226) INFO: 9epoch:train:6601-6700batch: iter_time=1.082e-04, loss_ctc=85.681, loss_att=74.500, acc=0.640, loss=77.811, grad_norm=3.638, loss_scale=1.000, learning_rate=1.377e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 09:19:06,920 (deepspeed_trainer:226) INFO: 9epoch:train:6701-6800batch: iter_time=1.112e-04, loss_ctc=78.381, loss_att=65.104, acc=0.643, loss=69.086, grad_norm=3.466, loss_scale=1.000, learning_rate=1.376e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:19:45,072 (deepspeed_trainer:226) INFO: 9epoch:train:6801-6900batch: iter_time=1.119e-04, loss_ctc=92.791, loss_att=73.026, acc=0.633, loss=78.972, grad_norm=3.898, loss_scale=1.000, learning_rate=1.375e-04, step_time=0.381 [2024-11-29 09:20:22,476] [INFO] [logging.py:129:log_dist] [Rank 0] step=127000, skipped=0, lr=[np.float64(0.00013746813811261736)], mom=[[0.9, 0.98]] [2024-11-29 09:20:22,477] [INFO] [timer.py:264:stop] epoch=0/micro_step=127000/global_step=127000, RunningAvgSamplesPerSec=43.049823835651104, CurrSamplesPerSec=45.57446478781075, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:20:22,478 (deepspeed_trainer:226) INFO: 9epoch:train:6901-7000batch: iter_time=1.051e-04, loss_ctc=87.577, loss_att=68.680, acc=0.630, loss=74.368, grad_norm=3.618, loss_scale=1.000, learning_rate=1.375e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:21:00,065 (deepspeed_trainer:226) INFO: 9epoch:train:7001-7100batch: iter_time=1.078e-04, loss_ctc=91.563, loss_att=77.857, acc=0.634, loss=81.956, grad_norm=4.021, loss_scale=1.000, learning_rate=1.374e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:21:37,639 (deepspeed_trainer:226) INFO: 9epoch:train:7101-7200batch: iter_time=1.115e-04, loss_ctc=95.249, loss_att=72.178, acc=0.651, loss=79.136, grad_norm=3.951, loss_scale=1.000, learning_rate=1.374e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:22:15,519 (deepspeed_trainer:226) INFO: 9epoch:train:7201-7300batch: iter_time=1.100e-04, loss_ctc=80.656, loss_att=67.171, acc=0.656, loss=71.212, grad_norm=3.642, loss_scale=1.000, learning_rate=1.373e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 09:22:52,962 (deepspeed_trainer:226) INFO: 9epoch:train:7301-7400batch: iter_time=1.125e-04, loss_ctc=95.639, loss_att=77.385, acc=0.635, loss=82.858, grad_norm=3.922, loss_scale=1.000, learning_rate=1.373e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:23:29,861 (deepspeed_trainer:226) INFO: 9epoch:train:7401-7500batch: iter_time=1.122e-04, loss_ctc=81.815, loss_att=67.760, acc=0.649, loss=71.951, grad_norm=3.508, loss_scale=1.000, learning_rate=1.372e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 09:23:33,449 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-008:0/16] 2024-11-29 09:24:00,980 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 09:24:06,950 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.1", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.1", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.1", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.1", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 09:24:06,950 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, [cnode7-008:0/16] 2024-11-29 09:24:06,952 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 09:24:22,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:33,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:23,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:23,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:23,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:34,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:34,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:35,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:24,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:35,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:24,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:24,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:24,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:36,253] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:36,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:24:36,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:10,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:11,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:22,398] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:11,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:23,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:12,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:23,270] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:23,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:13,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:13,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:13,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:13,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:24,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:25,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:25,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:27,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:59,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:59,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:25:59,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:00,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:01,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:12,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:01,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:02,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:02,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:12,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:13,375] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:13,478] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:14,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:15,375] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:15,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:19,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:46,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:47,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:47,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:48,249] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:48,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:49,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:00,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:00,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:50,483] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:26:50,519] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:01,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:01,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:02,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:03,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:03,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:27:09,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 09:28:15,502 (deepspeed_trainer:226) INFO: 9epoch:train:7501-7600batch: iter_time=2.417, loss_ctc=85.741, loss_att=61.712, acc=0.643, loss=68.923, grad_norm=3.698, loss_scale=1.000, learning_rate=1.372e-04, step_time=0.439 [cnode7-008:0/16] 2024-11-29 09:28:52,702 (deepspeed_trainer:226) INFO: 9epoch:train:7601-7700batch: iter_time=1.082e-04, loss_ctc=81.434, loss_att=65.366, acc=0.630, loss=70.165, grad_norm=3.653, loss_scale=1.000, learning_rate=1.371e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:29:30,480 (deepspeed_trainer:226) INFO: 9epoch:train:7701-7800batch: iter_time=1.208e-04, loss_ctc=82.841, loss_att=63.879, acc=0.642, loss=69.575, grad_norm=3.487, loss_scale=1.000, learning_rate=1.371e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 09:30:07,508 (deepspeed_trainer:226) INFO: 9epoch:train:7801-7900batch: iter_time=1.204e-04, loss_ctc=82.325, loss_att=66.352, acc=0.647, loss=71.133, grad_norm=3.540, loss_scale=1.000, learning_rate=1.370e-04, step_time=0.370 [2024-11-29 09:30:44,689] [INFO] [logging.py:129:log_dist] [Rank 0] step=128000, skipped=0, lr=[np.float64(0.00013693010449411554)], mom=[[0.9, 0.98]] [2024-11-29 09:30:44,690] [INFO] [timer.py:264:stop] epoch=0/micro_step=128000/global_step=128000, RunningAvgSamplesPerSec=43.0483961205211, CurrSamplesPerSec=44.981054991635766, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:30:44,691 (deepspeed_trainer:226) INFO: 9epoch:train:7901-8000batch: iter_time=1.133e-04, loss_ctc=82.175, loss_att=76.333, acc=0.605, loss=78.061, grad_norm=3.568, loss_scale=1.000, learning_rate=1.370e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:31:22,155 (deepspeed_trainer:226) INFO: 9epoch:train:8001-8100batch: iter_time=1.146e-04, loss_ctc=85.742, loss_att=64.998, acc=0.628, loss=71.203, grad_norm=3.951, loss_scale=1.000, learning_rate=1.369e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:31:59,608 (deepspeed_trainer:226) INFO: 9epoch:train:8101-8200batch: iter_time=1.190e-04, loss_ctc=89.343, loss_att=70.096, acc=0.648, loss=75.868, grad_norm=3.850, loss_scale=1.000, learning_rate=1.368e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:32:37,137 (deepspeed_trainer:226) INFO: 9epoch:train:8201-8300batch: iter_time=1.186e-04, loss_ctc=104.556, loss_att=71.357, acc=0.646, loss=81.301, grad_norm=5.294, loss_scale=1.000, learning_rate=1.368e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:33:14,480 (deepspeed_trainer:226) INFO: 9epoch:train:8301-8400batch: iter_time=1.144e-04, loss_ctc=92.562, loss_att=69.514, acc=0.633, loss=76.441, grad_norm=3.778, loss_scale=1.000, learning_rate=1.367e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 09:33:51,766 (deepspeed_trainer:226) INFO: 9epoch:train:8401-8500batch: iter_time=1.121e-04, loss_ctc=81.448, loss_att=65.714, acc=0.652, loss=70.440, grad_norm=3.475, loss_scale=1.000, learning_rate=1.367e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:34:28,567 (deepspeed_trainer:226) INFO: 9epoch:train:8501-8600batch: iter_time=1.137e-04, loss_ctc=82.314, loss_att=68.680, acc=0.634, loss=72.778, grad_norm=3.631, loss_scale=1.000, learning_rate=1.366e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 09:35:05,859 (deepspeed_trainer:226) INFO: 9epoch:train:8601-8700batch: iter_time=1.111e-04, loss_ctc=81.480, loss_att=67.941, acc=0.642, loss=71.972, grad_norm=3.365, loss_scale=1.000, learning_rate=1.366e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 09:35:43,166 (deepspeed_trainer:226) INFO: 9epoch:train:8701-8800batch: iter_time=1.086e-04, loss_ctc=87.992, loss_att=69.544, acc=0.627, loss=75.066, grad_norm=3.649, loss_scale=1.000, learning_rate=1.365e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 09:36:20,809 (deepspeed_trainer:226) INFO: 9epoch:train:8801-8900batch: iter_time=1.116e-04, loss_ctc=91.621, loss_att=73.319, acc=0.633, loss=78.828, grad_norm=3.872, loss_scale=1.000, learning_rate=1.365e-04, step_time=0.376 [2024-11-29 09:36:58,371] [INFO] [logging.py:129:log_dist] [Rank 0] step=129000, skipped=0, lr=[np.float64(0.00013639833921938588)], mom=[[0.9, 0.98]] [2024-11-29 09:36:58,372] [INFO] [timer.py:264:stop] epoch=0/micro_step=129000/global_step=129000, RunningAvgSamplesPerSec=43.05322351443164, CurrSamplesPerSec=39.958401253208045, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:36:58,373 (deepspeed_trainer:226) INFO: 9epoch:train:8901-9000batch: iter_time=1.105e-04, loss_ctc=87.971, loss_att=69.860, acc=0.645, loss=75.297, grad_norm=3.943, loss_scale=1.000, learning_rate=1.364e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:37:35,066 (deepspeed_trainer:226) INFO: 9epoch:train:9001-9100batch: iter_time=1.136e-04, loss_ctc=94.021, loss_att=76.476, acc=0.638, loss=81.729, grad_norm=3.738, loss_scale=1.000, learning_rate=1.364e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 09:38:13,014 (deepspeed_trainer:226) INFO: 9epoch:train:9101-9200batch: iter_time=1.141e-04, loss_ctc=84.940, loss_att=66.100, acc=0.654, loss=71.773, grad_norm=3.921, loss_scale=1.000, learning_rate=1.363e-04, step_time=0.379 [cnode7-008:0/16] 2024-11-29 09:38:49,791 (deepspeed_trainer:226) INFO: 9epoch:train:9201-9300batch: iter_time=1.140e-04, loss_ctc=85.614, loss_att=70.789, acc=0.633, loss=75.243, grad_norm=3.570, loss_scale=1.000, learning_rate=1.363e-04, step_time=0.367 [cnode7-008:0/16] 2024-11-29 09:39:21,351 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-008:0/16] 2024-11-29 09:39:49,020 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 09:39:55,000 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.3", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.3", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.3", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.3", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 09:39:55,000 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, [cnode7-008:0/16] 2024-11-29 09:39:55,003 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 09:40:09,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:21,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:10,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:10,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:21,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:22,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:11,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:11,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:11,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:22,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:11,768] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:22,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:11,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:23,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:23,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:23,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:57,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:10,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:59,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:59,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:59,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:40:59,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:11,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:00,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:00,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:12,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:12,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:12,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:12,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:13,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:13,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:02,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:45,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:46,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:47,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:59,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:48,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:48,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:48,501] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:48,870] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:01,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:01,557] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:01,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:02,029] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:02,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:02,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:03,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:41:53,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:31,197] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:34,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:35,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:35,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:36,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:47,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:36,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:36,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:48,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:49,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:50,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:51,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:51,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:51,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:52,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:42:42,687] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 09:43:29,422 (deepspeed_trainer:226) INFO: 9epoch:train:9301-9400batch: iter_time=2.423, loss_ctc=82.788, loss_att=65.516, acc=0.654, loss=70.695, grad_norm=3.916, loss_scale=1.000, learning_rate=1.362e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 09:44:06,939 (deepspeed_trainer:226) INFO: 9epoch:train:9401-9500batch: iter_time=1.197e-04, loss_ctc=84.224, loss_att=61.700, acc=0.643, loss=68.450, grad_norm=3.934, loss_scale=1.000, learning_rate=1.362e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:44:43,776 (deepspeed_trainer:226) INFO: 9epoch:train:9501-9600batch: iter_time=1.134e-04, loss_ctc=82.282, loss_att=64.089, acc=0.640, loss=69.554, grad_norm=3.896, loss_scale=1.000, learning_rate=1.361e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 09:45:20,883 (deepspeed_trainer:226) INFO: 9epoch:train:9601-9700batch: iter_time=1.110e-04, loss_ctc=83.584, loss_att=64.315, acc=0.629, loss=70.076, grad_norm=3.968, loss_scale=1.000, learning_rate=1.361e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 09:45:58,679 (deepspeed_trainer:226) INFO: 9epoch:train:9701-9800batch: iter_time=1.143e-04, loss_ctc=84.102, loss_att=69.396, acc=0.645, loss=73.810, grad_norm=3.630, loss_scale=1.000, learning_rate=1.360e-04, step_time=0.377 [cnode7-008:0/16] 2024-11-29 09:46:36,288 (deepspeed_trainer:226) INFO: 9epoch:train:9801-9900batch: iter_time=1.093e-04, loss_ctc=80.702, loss_att=68.684, acc=0.618, loss=72.272, grad_norm=4.027, loss_scale=1.000, learning_rate=1.360e-04, step_time=0.376 [2024-11-29 09:47:13,896] [INFO] [logging.py:129:log_dist] [Rank 0] step=130000, skipped=0, lr=[np.float64(0.00013587272151096603)], mom=[[0.9, 0.98]] [2024-11-29 09:47:13,897] [INFO] [timer.py:264:stop] epoch=0/micro_step=130000/global_step=130000, RunningAvgSamplesPerSec=43.05843972552318, CurrSamplesPerSec=42.48552184765042, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:47:13,898 (deepspeed_trainer:226) INFO: 9epoch:train:9901-10000batch: iter_time=1.248e-04, loss_ctc=84.979, loss_att=63.757, acc=0.638, loss=70.115, grad_norm=4.098, loss_scale=1.000, learning_rate=1.359e-04, step_time=0.376 [cnode7-008:0/16] 2024-11-29 09:47:50,775 (deepspeed_trainer:226) INFO: 9epoch:train:10001-10100batch: iter_time=1.160e-04, loss_ctc=89.485, loss_att=68.123, acc=0.645, loss=74.545, grad_norm=4.194, loss_scale=1.000, learning_rate=1.358e-04, step_time=0.369 [cnode7-008:0/16] 2024-11-29 09:48:28,806 (deepspeed_trainer:226) INFO: 9epoch:train:10101-10200batch: iter_time=1.123e-04, loss_ctc=103.585, loss_att=73.673, acc=0.650, loss=82.679, grad_norm=4.859, loss_scale=1.000, learning_rate=1.358e-04, step_time=0.380 [cnode7-008:0/16] 2024-11-29 09:49:06,235 (deepspeed_trainer:226) INFO: 9epoch:train:10201-10300batch: iter_time=1.135e-04, loss_ctc=86.340, loss_att=62.856, acc=0.638, loss=69.886, grad_norm=3.992, loss_scale=1.000, learning_rate=1.357e-04, step_time=0.374 [cnode7-008:0/16] 2024-11-29 09:49:43,078 (deepspeed_trainer:226) INFO: 9epoch:train:10301-10400batch: iter_time=1.136e-04, loss_ctc=77.835, loss_att=65.389, acc=0.654, loss=69.121, grad_norm=3.420, loss_scale=1.000, learning_rate=1.357e-04, step_time=0.368 [cnode7-008:0/16] 2024-11-29 09:50:20,432 (deepspeed_trainer:226) INFO: 9epoch:train:10401-10500batch: iter_time=1.130e-04, loss_ctc=83.368, loss_att=70.342, acc=0.638, loss=74.252, grad_norm=3.312, loss_scale=1.000, learning_rate=1.356e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 09:50:58,241 (deepspeed_trainer:226) INFO: 9epoch:train:10501-10600batch: iter_time=1.081e-04, loss_ctc=86.914, loss_att=68.971, acc=0.632, loss=74.369, grad_norm=3.849, loss_scale=1.000, learning_rate=1.356e-04, step_time=0.378 [cnode7-008:0/16] 2024-11-29 09:51:36,397 (deepspeed_trainer:226) INFO: 9epoch:train:10601-10700batch: iter_time=1.105e-04, loss_ctc=88.557, loss_att=69.866, acc=0.638, loss=75.446, grad_norm=4.037, loss_scale=1.000, learning_rate=1.355e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 09:52:13,922 (deepspeed_trainer:226) INFO: 9epoch:train:10701-10800batch: iter_time=1.146e-04, loss_ctc=85.163, loss_att=71.441, acc=0.635, loss=75.564, grad_norm=3.782, loss_scale=1.000, learning_rate=1.355e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:52:51,283 (deepspeed_trainer:226) INFO: 9epoch:train:10801-10900batch: iter_time=1.136e-04, loss_ctc=92.949, loss_att=69.457, acc=0.654, loss=76.444, grad_norm=4.254, loss_scale=1.000, learning_rate=1.354e-04, step_time=0.373 [2024-11-29 09:53:28,408] [INFO] [logging.py:129:log_dist] [Rank 0] step=131000, skipped=0, lr=[np.float64(0.00013535313382445377)], mom=[[0.9, 0.98]] [2024-11-29 09:53:28,409] [INFO] [timer.py:264:stop] epoch=0/micro_step=131000/global_step=131000, RunningAvgSamplesPerSec=43.06244373848898, CurrSamplesPerSec=42.73646397405208, MemAllocated=2.04GB, MaxMemAllocated=23.06GB [cnode7-008:0/16] 2024-11-29 09:53:28,410 (deepspeed_trainer:226) INFO: 9epoch:train:10901-11000batch: iter_time=1.175e-04, loss_ctc=87.916, loss_att=70.299, acc=0.647, loss=75.559, grad_norm=3.848, loss_scale=1.000, learning_rate=1.354e-04, step_time=0.371 [cnode7-008:0/16] 2024-11-29 09:54:05,960 (deepspeed_trainer:226) INFO: 9epoch:train:11001-11100batch: iter_time=1.119e-04, loss_ctc=87.793, loss_att=67.891, acc=0.646, loss=73.872, grad_norm=3.895, loss_scale=1.000, learning_rate=1.353e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 09:54:44,061 (deepspeed_trainer:226) INFO: 9epoch:train:11101-11200batch: iter_time=1.126e-04, loss_ctc=86.142, loss_att=70.849, acc=0.640, loss=75.454, grad_norm=4.166, loss_scale=1.000, learning_rate=1.353e-04, step_time=0.381 [cnode7-008:0/16] 2024-11-29 09:55:06,535 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-008:0/16] 2024-11-29 09:55:34,083 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-008:0/16] 2024-11-29 09:55:39,678 (abs_task:1807) INFO: [train] dataset: ESPnetDataset( speech: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/wav.scp/split.7", "type": "kaldi_ark"} text_prev: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.prev/split.7", "type": "text"} text_ctc: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text.ctc/split.7", "type": "text"} text: {"path": "exp_owsm/s2t_stats_raw_bpe50000/splits8/text/split.7", "type": "text"} preprocess: ) [cnode7-008:0/16] 2024-11-29 09:55:39,678 (abs_task:1808) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=28521, batch_size=256, key_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, [cnode7-008:0/16] 2024-11-29 09:55:39,681 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-11-29 09:55:54,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:54,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:54,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:54,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:56,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:56,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:56,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:55:56,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:08,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:08,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:08,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:08,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:09,150] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:09,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:09,737] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:09,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:42,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:42,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:43,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:43,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:43,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:44,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:44,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:44,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:56,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:57,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:58,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:58,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:58,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:58,855] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:56:59,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:01,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:31,409] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:32,129] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:32,502] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:32,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:32,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:33,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:33,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:34,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:46,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:46,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:47,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:48,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:48,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:48,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:51,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:57:51,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:18,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:20,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:20,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:20,943] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:21,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:21,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:22,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:22,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:35,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:35,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:36,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:37,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:37,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:37,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:40,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-11-29 09:58:43,503] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-008:0/16] 2024-11-29 09:59:29,310 (deepspeed_trainer:226) INFO: 9epoch:train:11201-11300batch: iter_time=2.410, loss_ctc=86.011, loss_att=65.588, acc=0.652, loss=71.687, grad_norm=4.103, loss_scale=1.000, learning_rate=1.352e-04, step_time=0.442 [cnode7-008:0/16] 2024-11-29 10:00:06,919 (deepspeed_trainer:226) INFO: 9epoch:train:11301-11400batch: iter_time=1.172e-04, loss_ctc=77.218, loss_att=56.170, acc=0.654, loss=62.491, grad_norm=3.596, loss_scale=1.000, learning_rate=1.352e-04, step_time=0.375 [cnode7-008:0/16] 2024-11-29 10:00:45,243 (deepspeed_trainer:226) INFO: 9epoch:train:11401-11500batch: iter_time=1.133e-04, loss_ctc=82.452, loss_att=65.542, acc=0.643, loss=70.600, grad_norm=4.034, loss_scale=1.000, learning_rate=1.351e-04, step_time=0.383 [cnode7-008:0/16] 2024-11-29 10:01:23,395 (deepspeed_trainer:226) INFO: 9epoch:train:11501-11600batch: iter_time=1.139e-04, loss_ctc=84.792, loss_att=67.828, acc=0.624, loss=72.896, grad_norm=4.278, loss_scale=1.000, learning_rate=1.351e-04, step_time=0.382 [cnode7-008:0/16] 2024-11-29 10:02:00,799 (deepspeed_trainer:226) INFO: 9epoch:train:11601-11700batch: iter_time=1.117e-04, loss_ctc=82.322, loss_att=63.105, acc=0.647, loss=68.882, grad_norm=3.643, loss_scale=1.000, learning_rate=1.350e-04, step_time=0.373 [cnode7-008:0/16] 2024-11-29 10:02:37,980 (deepspeed_trainer:226) INFO: 9epoch:train:11701-11800batch: iter_time=1.126e-04, loss_ctc=84.639, loss_att=73.151, acc=0.617, loss=76.576, grad_norm=3.944, loss_scale=1.000, learning_rate=1.350e-04, step_time=0.372 [cnode7-008:0/16] 2024-11-29 10:03:15,701 (deepspeed_trainer:226) INFO: 9epoch:train:11801-11900batch: iter_time=1.132e-04, loss_ctc=85.738, loss_att=64.125, acc=0.647, loss=70.596, grad_norm=3.984, loss_scale=1.000, learning_rate=1.349e-04, step_time=0.376 Process SpawnProcess-5: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 204, in train_one_epoch loss, stats, weight = model(**batch) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/utils/nvtx.py", line 18, in wrapped_fn ret_val = func(*args, **kwargs) File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/engine.py", line 1899, in forward loss = self.module(*inputs, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py", line 225, in forward loss_att, acc_att, cer_att, wer_att = self._calc_att_loss( File "/mnt/home/williamchen/espnet/espnet2/s2t/espnet_model.py", line 391, in _calc_att_loss decoder_out, _ = self.decoder( File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/espnet/espnet2/asr/decoder/transformer_decoder.py", line 155, in forward x, tgt_mask, memory, memory_mask = checkpoint(decoder_layer, File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/_dynamo/eval_frame.py", line 600, in _fn return fn(*args, **kwargs) File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/activation_checkpointing/checkpointing.py", line 951, in checkpoint CheckpointFunction.apply(function, all_outputs, *args) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/autograd/function.py", line 574, in apply return super().apply(*args, **kwargs) # type: ignore[misc] File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/activation_checkpointing/checkpointing.py", line 542, in forward outputs = run_function(*inputs_cuda) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/espnet/espnet/nets/pytorch_backend/transformer/decoder_layer.py", line 170, in forward x = residual + self.dropout(self.feed_forward(x)) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/espnet/espnet/nets/pytorch_backend/transformer/positionwise_feed_forward.py", line 32, in forward return self.w_2(self.dropout(self.activation(self.w_1(x)))) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl return forward_call(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 117, in forward return F.linear(input, self.weight, self.bias) torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 6.91 GiB. GPU 4 has a total capacity of 79.11 GiB of which 4.91 GiB is free. Including non-PyTorch memory, this process has 74.19 GiB memory in use. Of the allocated memory 67.05 GiB is allocated by PyTorch, and 1.11 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) cnode7-008:1507505:1516912 [0] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507505:1508447 [0] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507510:1516903 [5] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507510:1508438 [5] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507511:1516902 [6] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507511:1508441 [6] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507512:1516908 [7] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507512:1508437 [7] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507507:1516898 [2] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507507:1508433 [2] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507508:1516900 [3] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507508:1508443 [3] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507506:1516904 [1] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-008:1507506:1508434 [1] NCCL INFO [Service thread] Connection closed by localRank 4 W1129 10:03:58.959000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507505 via signal SIGTERM W1129 10:03:58.960000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507506 via signal SIGTERM W1129 10:03:58.960000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507507 via signal SIGTERM W1129 10:03:58.961000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507508 via signal SIGTERM W1129 10:03:58.962000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507510 via signal SIGTERM W1129 10:03:58.962000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507511 via signal SIGTERM W1129 10:03:58.962000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 1507512 via signal SIGTERM Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/runpy.py", line 196, in _run_module_as_main return _run_code(code, main_globals, None, File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/runpy.py", line 86, in _run_code exec(code, run_globals) File "/mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py", line 23, in main() File "/mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py", line 19, in main S2TTask.main(cmd=cmd) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1273, in main while not ProcessContext(processes, error_files).join(): File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 178, in join raise ProcessExitedException( torch.multiprocessing.spawn.ProcessExitedException: process 4 terminated with exit code 1 srun: error: cnode7-008: task 0: Exited with exit code 1 [rank12]:[E1129 10:13:38.545817383 ProcessGroupNCCL.cpp:607] [Rank 12] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600001 milliseconds before timing out. [rank13]:[E1129 10:13:38.545825683 ProcessGroupNCCL.cpp:607] [Rank 13] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600005 milliseconds before timing out. [rank14]:[E1129 10:13:38.545830810 ProcessGroupNCCL.cpp:607] [Rank 14] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600009 milliseconds before timing out. [rank11]:[E1129 10:13:38.545839780 ProcessGroupNCCL.cpp:607] [Rank 11] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600001 milliseconds before timing out. [rank15]:[E1129 10:13:38.545848677 ProcessGroupNCCL.cpp:607] [Rank 15] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600009 milliseconds before timing out. [rank9]:[E1129 10:13:38.545840383 ProcessGroupNCCL.cpp:607] [Rank 9] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600003 milliseconds before timing out. [rank10]:[E1129 10:13:38.545842973 ProcessGroupNCCL.cpp:607] [Rank 10] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600004 milliseconds before timing out. [rank11]:[E1129 10:13:38.545867114 ProcessGroupNCCL.cpp:607] [Rank 11] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600014 milliseconds before timing out. [rank8]:[E1129 10:13:38.545869817 ProcessGroupNCCL.cpp:607] [Rank 8] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600005 milliseconds before timing out. [rank9]:[E1129 10:13:38.545944686 ProcessGroupNCCL.cpp:607] [Rank 9] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600004 milliseconds before timing out. [rank11]:[E1129 10:13:38.554708564 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 11] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank11]:[E1129 10:13:38.554719488 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 11] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank11]:[E1129 10:13:38.554723059 ProcessGroupNCCL.cpp:621] [Rank 11] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank9]:[E1129 10:13:38.554714926 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 9] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank9]:[E1129 10:13:38.554726367 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 9] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank9]:[E1129 10:13:38.554731756 ProcessGroupNCCL.cpp:621] [Rank 9] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank9]:[E1129 10:13:38.556502660 ProcessGroupNCCL.cpp:670] [Rank 9] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank8]:[E1129 10:13:38.556515575 ProcessGroupNCCL.cpp:670] [Rank 8] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank12]:[E1129 10:13:38.556524889 ProcessGroupNCCL.cpp:670] [Rank 12] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank13]:[E1129 10:13:38.556525452 ProcessGroupNCCL.cpp:670] [Rank 13] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank11]:[E1129 10:13:38.556528261 ProcessGroupNCCL.cpp:670] [Rank 11] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank14]:[E1129 10:13:38.556536222 ProcessGroupNCCL.cpp:670] [Rank 14] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank10]:[E1129 10:13:38.556536176 ProcessGroupNCCL.cpp:670] [Rank 10] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). [rank15]:[E1129 10:13:38.556539303 ProcessGroupNCCL.cpp:670] [Rank 15] Work WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) timed out in blocking wait (TORCH_NCCL_BLOCKING_WAIT=1). cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 0 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637491:647077 [0] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637492:647084 [1] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637496:647083 [5] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637494:647082 [3] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637497:647078 [6] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-009:637498:647081 [7] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637493:647090 [2] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637495:647080 [4] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-009:637495:637495 [4] NCCL INFO comm 0x5555c2c27fe0 rank 12 nranks 16 cudaDev 4 busId 9d000 - Abort COMPLETE [rank12]:[E1129 10:13:38.921590139 ProcessGroupNCCL.cpp:621] [Rank 12] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank12]:[E1129 10:13:38.921605651 ProcessGroupNCCL.cpp:627] [Rank 12] To avoid data inconsistency, we are taking the entire process down. [rank12]:[E1129 10:13:38.927609405 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 12] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank12]:[E1129 10:13:38.927623764 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 12] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank12]:[E1129 10:13:38.927628329 ProcessGroupNCCL.cpp:621] [Rank 12] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637491:637491 [0] NCCL INFO comm 0x5555b8bd9f40 rank 8 nranks 16 cudaDev 0 busId 1b000 - Abort COMPLETE [rank8]:[E1129 10:13:38.975597122 ProcessGroupNCCL.cpp:621] [Rank 8] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank8]:[E1129 10:13:38.975614122 ProcessGroupNCCL.cpp:627] [Rank 8] To avoid data inconsistency, we are taking the entire process down. [rank8]:[E1129 10:13:38.977312704 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 8] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank8]:[E1129 10:13:38.977325762 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 8] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank8]:[E1129 10:13:38.977331449 ProcessGroupNCCL.cpp:621] [Rank 8] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637493:637493 [2] NCCL INFO comm 0x5555c2c0bbd0 rank 10 nranks 16 cudaDev 2 busId 52000 - Abort COMPLETE [rank10]:[E1129 10:13:38.001905402 ProcessGroupNCCL.cpp:621] [Rank 10] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank10]:[E1129 10:13:38.001922011 ProcessGroupNCCL.cpp:627] [Rank 10] To avoid data inconsistency, we are taking the entire process down. [rank10]:[E1129 10:13:38.003804531 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 10] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank10]:[E1129 10:13:38.003818325 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 10] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank10]:[E1129 10:13:38.003823069 ProcessGroupNCCL.cpp:621] [Rank 10] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637494:637494 [3] NCCL INFO comm 0x5555c2bfa560 rank 11 nranks 16 cudaDev 3 busId 61000 - Abort COMPLETE [rank11]:[E1129 10:13:38.006248838 ProcessGroupNCCL.cpp:621] [Rank 11] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank11]:[E1129 10:13:38.006262879 ProcessGroupNCCL.cpp:627] [Rank 11] To avoid data inconsistency, we are taking the entire process down. cnode7-009:637496:637496 [5] NCCL INFO comm 0x5555c2bf8710 rank 13 nranks 16 cudaDev 5 busId c3000 - Abort COMPLETE [rank13]:[E1129 10:13:39.065655540 ProcessGroupNCCL.cpp:621] [Rank 13] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank13]:[E1129 10:13:39.065676239 ProcessGroupNCCL.cpp:627] [Rank 13] To avoid data inconsistency, we are taking the entire process down. [rank13]:[E1129 10:13:39.067749033 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 13] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank13]:[E1129 10:13:39.067763906 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 13] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank13]:[E1129 10:13:39.067768600 ProcessGroupNCCL.cpp:621] [Rank 13] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637498:637498 [7] NCCL INFO comm 0x5555d6bd2ca0 rank 15 nranks 16 cudaDev 7 busId df000 - Abort COMPLETE [rank15]:[E1129 10:13:39.080810598 ProcessGroupNCCL.cpp:621] [Rank 15] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank15]:[E1129 10:13:39.080826570 ProcessGroupNCCL.cpp:627] [Rank 15] To avoid data inconsistency, we are taking the entire process down. [rank15]:[E1129 10:13:39.082743599 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 15] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank15]:[E1129 10:13:39.082758749 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 15] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank15]:[E1129 10:13:39.082763309 ProcessGroupNCCL.cpp:621] [Rank 15] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637497:637497 [6] NCCL INFO comm 0x5555d6c24420 rank 14 nranks 16 cudaDev 6 busId d1000 - Abort COMPLETE [rank14]:[E1129 10:13:39.121885320 ProcessGroupNCCL.cpp:621] [Rank 14] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank14]:[E1129 10:13:39.121905159 ProcessGroupNCCL.cpp:627] [Rank 14] To avoid data inconsistency, we are taking the entire process down. [rank14]:[E1129 10:13:39.123944821 ProcessGroupNCCL.cpp:1664] [PG 0 (default_pg) Rank 14] Exception (either an error or timeout) detected by watchdog at work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank14]:[E1129 10:13:39.123960327 ProcessGroupNCCL.cpp:1709] [PG 0 (default_pg) Rank 14] Timeout at NCCL work: 660070, last enqueued NCCL work: 660070, last completed NCCL work: 660069. [rank14]:[E1129 10:13:39.123965069 ProcessGroupNCCL.cpp:621] [Rank 14] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. cnode7-009:637492:637492 [1] NCCL INFO comm 0x5555c2bf3600 rank 9 nranks 16 cudaDev 1 busId 43000 - Abort COMPLETE [rank9]:[E1129 10:13:39.136261586 ProcessGroupNCCL.cpp:621] [Rank 9] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, subsequent GPU operations might run on corrupted/incomplete data. [rank9]:[E1129 10:13:39.136283260 ProcessGroupNCCL.cpp:627] [Rank 9] To avoid data inconsistency, we are taking the entire process down. Process SpawnProcess-1: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 8] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600005 milliseconds before timing out. Process SpawnProcess-5: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 12] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600001 milliseconds before timing out. Process SpawnProcess-4: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 11] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600001 milliseconds before timing out. Process SpawnProcess-2: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 9] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600003 milliseconds before timing out. Process SpawnProcess-7: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 14] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600009 milliseconds before timing out. Process SpawnProcess-3: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 10] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600004 milliseconds before timing out. Process SpawnProcess-6: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 13] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600005 milliseconds before timing out. Process SpawnProcess-8: Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap self.run() File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1568, in main_worker cls.trainer.run( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 132, in run cls.train_one_epoch( File "/mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py", line 208, in train_one_epoch stats, weight = recursive_average(stats, weight, True) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 53, in recursive_average obj = recursive_sum(obj, weight, distributed) File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in recursive_sum return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 14, in return {k: recursive_sum(v, weight, distributed) for k, v in obj.items()} File "/mnt/home/williamchen/espnet/espnet2/torch_utils/recursive_op.py", line 22, in recursive_sum torch.distributed.all_gather(lst, obj) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 79, in wrapper return func(*args, **kwargs) File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3113, in all_gather work.wait() torch.distributed.DistBackendError: [Rank 15] Watchdog caught collective operation timeout: WorkNCCL(SeqNum=660070, OpType=ALLGATHER, NumelIn=1, NumelOut=16, Timeout(ms)=600000) ran for 600009 milliseconds before timing out. W1129 10:13:49.075000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 637493 via signal SIGTERM W1129 10:13:49.076000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 637496 via signal SIGTERM W1129 10:13:49.076000 23456244418368 torch/multiprocessing/spawn.py:146] Terminating process 637498 via signal SIGTERM Traceback (most recent call last): File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/runpy.py", line 196, in _run_module_as_main return _run_code(code, main_globals, None, File "/mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/runpy.py", line 86, in _run_code exec(code, run_globals) File "/mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py", line 23, in main() File "/mnt/home/williamchen/espnet/espnet2/bin/s2t_train.py", line 19, in main S2TTask.main(cmd=cmd) File "/mnt/home/williamchen/espnet/espnet2/tasks/abs_task.py", line 1273, in main while not ProcessContext(processes, error_files).join(): File "/mnt/home/williamchen/.local/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 178, in join raise ProcessExitedException( torch.multiprocessing.spawn.ProcessExitedException: process 6 terminated with exit code 1 srun: error: cnode7-009: task 1: Exited with exit code 1 # Accounting: begin_time=1732778191 # Accounting: end_time=1732846442 # Accounting: time=68251 threads=1 # Finished at Fri Nov 29 10:14:02 CST 2024 with status 1