# Running on cnode7-018 # Started at Tue Dec 10 17:28:25 CST 2024 # SLURMD_NODENAME=cnode7-018 # 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=6579 # SLURM_JOB_CPUS_PER_NODE='224(x2)' # SLURM_JOB_END_TIME=1765358903 # SLURM_JOB_GID=1026 # SLURM_JOB_GPUS=0,1,2,3,4,5,6,7 # SLURM_JOB_ID=6579 # SLURM_JOB_NAME=exp_owsm/s2t_train_05b_ds_raw_bpe50000/train.log # SLURM_JOB_NODELIST='cnode7-[018-019]' # SLURM_JOB_NUM_NODES=2 # SLURM_JOB_PARTITION=sa # SLURM_JOB_QOS=normal # SLURM_JOB_START_TIME=1733822903 # 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-[018-019]' # 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=1405965 # SLURM_TOPOLOGY_ADDR=cnode7-018 # 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_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --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_8efcdeaa-a3e1-4bf9-aba4-b79a5f252245 [2024-12-10 17:28:40,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:28:40,486] [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_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --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_8efcdeaa-a3e1-4bf9-aba4-b79a5f252245 /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_data_path_and_name_and_type dump/raw/dev_v3/text.ctc,text_ctc,text --valid_data_path_and_name_and_type dump/raw/dev_v3/text,text,text --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_8efcdeaa-a3e1-4bf9-aba4-b79a5f252245 [2024-12-10 17:29:01,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:01,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:02,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,275] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:29:03,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [W1210 17:29:09.459821998 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:09.899825985 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.308379466 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.600885706 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.644116236 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.809050863 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.825514184 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.904649693 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.867370348 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:10.294573090 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:11.669143636 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:11.676328438 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [cnode7-018:0/16] 2024-12-10 17:29:11,119 (s2t:462) INFO: Vocabulary size: 50002 [W1210 17:29:11.927462507 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:11.988470625 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:11.138693975 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [W1210 17:29:11.271814089 Utils.hpp:164] Warning: Environment variable NCCL_BLOCKING_WAIT is deprecated; use TORCH_NCCL_BLOCKING_WAIT instead (function operator()) [cnode7-018:0/16] 2024-12-10 17:29:15,068 (abs_task:1383) INFO: pytorch.version=2.4.0+cu121, cuda.available=True, cudnn.version=90100, cudnn.benchmark=False, cudnn.deterministic=True [cnode7-018:0/16] 2024-12-10 17:29:15,074 (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 [cnode7-018:0/16] 2024-12-10 17:29:15,074 (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 ) [cnode7-018:0/16] 2024-12-10 17:29:15,074 (abs_task:1388) INFO: Scheduler: None [cnode7-018:0/16] 2024-12-10 17:29:15,079 (abs_task:1397) INFO: Saving the configuration in exp_owsm/s2t_train_05b_ds_raw_bpe50000/config.yaml [cnode7-018:0/16] 2024-12-10 17:29:19,643 (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:240: 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( [2024-12-10 17:29:20,882] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:20,886] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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:240: 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( [2024-12-10 17:29:21,133] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,134] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-12-10 17:29:21,246] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,248] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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:240: 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:240: 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( [2024-12-10 17:29:21,574] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,575] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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( [2024-12-10 17:29:21,621] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,622] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-12-10 17:29:21,627] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,628] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-12-10 17:29:21,760] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:21,761] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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( [2024-12-10 17:29:22,239] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,242] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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:240: 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( [2024-12-10 17:29:22,870] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,871] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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:240: 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( [cnode7-018:0/16] 2024-12-10 17:29:22,582 (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: ) [cnode7-018:0/16] 2024-12-10 17:29:22,582 (abs_task:1808) INFO: [valid] Batch sampler: SortedBatchSampler(N-batch=74743, batch_size=16, shape_file=exp_owsm/s2t_stats_raw_bpe50000/valid/speech_shape, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 17:29:22,588 (abs_task:1809) INFO: [valid] mini-batch sizes summary: N-batch=74743, mean=16.0, min=16, max=17 [2024-12-10 17:29:22,595] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,596] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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:240: 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( [cnode7-018:0/16] 2024-12-10 17:29:22,644 (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 ================================================================= [2024-12-10 17:29:22,644] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,645] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed info: version=0.15.3, git-hash=unknown, git-branch=unknown [2024-12-10 17:29:22,645] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-018:1407096:1407096 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407096:1407096 [0] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407096:1407096 [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 cnode7-018:1407096:1407096 [0] NCCL INFO cudaDriverVersion 12020 NCCL version 2.20.5+cuda12.4 [2024-12-10 17:29:22,746] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,747] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 [2024-12-10 17:29:22,758] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,759] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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( [2024-12-10 17:29:22,790] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,792] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-018:1407097:1407097 [1] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407097:1407097 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407097:1407097 [1] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407097:1407097 [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 cnode7-019:1739576:1739576 [2] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739576:1739576 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739576:1739576 [2] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739581:1739581 [7] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739581:1739581 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739581:1739581 [7] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739581:1739581 [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 cnode7-019:1739580:1739580 [6] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739580:1739580 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739580:1739580 [6] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739580:1739580 [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 cnode7-018:1407100:1407100 [4] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407100:1407100 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407100:1407100 [4] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407100:1407100 [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 cnode7-019:1739579:1739579 [5] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739579:1739579 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739579:1739579 [5] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739579:1739579 [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 cnode7-018:1407098:1407098 [2] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407098:1407098 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407098:1407098 [2] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407098:1407098 [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 cnode7-019:1739578:1739578 [4] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739578:1739578 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739578:1739578 [4] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739578:1739578 [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 cnode7-019:1739575:1739575 [1] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739575:1739575 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739575:1739575 [1] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739575:1739575 [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 cnode7-018:1407102:1407102 [6] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407102:1407102 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407102:1407102 [6] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407102:1407102 [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 cnode7-019:1739577:1739577 [3] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739577:1739577 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739577:1739577 [3] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739577:1739577 [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 /mnt/home/williamchen/espnet/espnet2/train/deepspeed_trainer.py:240: 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( cnode7-018:1407103:1407103 [7] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407103:1407103 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407103:1407103 [7] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407103:1407103 [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 cnode7-019:1739574:1739574 [0] NCCL INFO cudaDriverVersion 12020 cnode7-019:1739574:1739574 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739574:1739574 [0] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.186<0> cnode7-019:1739574:1739574 [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 cnode7-019:1739576:1739576 [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 [2024-12-10 17:29:22,951] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:22,952] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-018:1407096:1407758 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407102:1407766 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407097:1407760 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407099:1407099 [3] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407099:1407099 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407099:1407099 [3] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407099:1407099 [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 cnode7-018:1407100:1407759 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407098:1407767 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407096:1407758 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407096:1407758 [0] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407096:1407758 [0] NCCL INFO Using network IB [2024-12-10 17:29:23,082] [INFO] [comm.py:652:init_distributed] cdb=None [2024-12-10 17:29:23,084] [INFO] [config.py:733:__init__] Config mesh_device None world_size = 16 cnode7-018:1407102:1407766 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407102:1407766 [6] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407102:1407766 [6] NCCL INFO Using network IB cnode7-018:1407097:1407760 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407097:1407760 [1] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407097:1407760 [1] NCCL INFO Using network IB cnode7-018:1407103:1407771 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739574:1740207 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739576:1740203 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739581:1740202 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739580:1740209 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407100:1407759 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407100:1407759 [4] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407100:1407759 [4] NCCL INFO Using network IB cnode7-019:1739579:1740205 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407098:1407767 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407098:1407767 [2] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407098:1407767 [2] NCCL INFO Using network IB cnode7-019:1739578:1740204 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739575:1740208 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-019:1739577:1740206 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407103:1407771 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407103:1407771 [7] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407103:1407771 [7] NCCL INFO Using network IB cnode7-019:1739574:1740207 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739574:1740207 [0] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739574:1740207 [0] NCCL INFO Using network IB cnode7-019:1739581:1740202 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739581:1740202 [7] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739581:1740202 [7] NCCL INFO Using network IB cnode7-019:1739580:1740209 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739580:1740209 [6] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739580:1740209 [6] NCCL INFO Using network IB cnode7-019:1739579:1740205 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739579:1740205 [5] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739579:1740205 [5] NCCL INFO Using network IB cnode7-019:1739578:1740204 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739578:1740204 [4] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739578:1740204 [4] NCCL INFO Using network IB cnode7-019:1739575:1740208 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739575:1740208 [1] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739575:1740208 [1] NCCL INFO Using network IB cnode7-019:1739577:1740206 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739577:1740206 [3] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739577:1740206 [3] NCCL INFO Using network IB cnode7-019:1739576:1740203 [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 enp170s0np0:10.225.1.186<0> cnode7-019:1739576:1740203 [2] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739576:1740203 [2] NCCL INFO Using network IB cnode7-018:1407101:1407101 [5] NCCL INFO cudaDriverVersion 12020 cnode7-018:1407101:1407101 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407101:1407101 [5] NCCL INFO Bootstrap : Using enp170s0np0:10.225.1.180<0> cnode7-018:1407101:1407101 [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 cnode7-018:1407099:1407783 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407099:1407783 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407099:1407783 [3] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407099:1407783 [3] NCCL INFO Using network IB cnode7-018:1407101:1407851 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to ^lo,docker,virbr,vmnet,vboxnet cnode7-018:1407101:1407851 [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 enp170s0np0:10.225.1.180<0> cnode7-018:1407101:1407851 [5] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407101:1407851 [5] NCCL INFO Using network IB cnode7-019:1739574:1740207 [0] NCCL INFO comm 0x555564db6160 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739576:1740203 [2] NCCL INFO comm 0x55556e619a10 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739581:1740202 [7] NCCL INFO comm 0x555580e1fe30 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407099:1407783 [3] NCCL INFO comm 0x55558141d780 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739580:1740209 [6] NCCL INFO comm 0x55557c420110 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407100:1407759 [4] NCCL INFO comm 0x55557c01bc00 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739579:1740205 [5] NCCL INFO comm 0x555560c1a8b0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407098:1407767 [2] NCCL INFO comm 0x555561a25360 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407101:1407851 [5] NCCL INFO comm 0x55557c41c200 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739578:1740204 [4] NCCL INFO comm 0x55557261e970 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739575:1740208 [1] NCCL INFO comm 0x555561a25020 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407096:1407758 [0] NCCL INFO comm 0x5555637cce40 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407102:1407766 [6] NCCL INFO comm 0x55557261e440 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xa4a394c9439e5d9f - Init START cnode7-019:1739577:1740206 [3] NCCL INFO comm 0x55558161bf60 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407097:1407760 [1] NCCL INFO comm 0x55556c41b400 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407103:1407771 [7] NCCL INFO comm 0x55557261d7b0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xa4a394c9439e5d9f - Init START cnode7-018:1407096:1407758 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407096:1407758 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-018:1407103:1407771 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-018:1407102:1407766 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-018:1407097:1407760 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407097:1407760 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-018:1407100:1407759 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-019:1739581:1740202 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739581:1740202 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-019:1739577:1740206 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739577:1740206 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-018:1407099:1407783 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407099:1407783 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-018:1407098:1407767 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407098:1407767 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-018:1407101:1407851 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-019:1739580:1740209 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739580:1740209 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-019:1739579:1740205 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739579:1740205 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-019:1739578:1740204 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739578:1740204 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-019:1739575:1740208 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739575:1740208 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-018:1407097:1407760 [1] NCCL INFO comm 0x55556c41b400 rank 1 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407097:1407760 [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 cnode7-018:1407097:1407760 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407096:1407758 [0] NCCL INFO comm 0x5555637cce40 rank 0 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 00/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 01/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 03/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 04/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 05/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 06/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 07/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 08/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 09/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 11/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 12/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 13/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 14/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-018:1407096:1407758 [0] NCCL INFO Channel 15/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-018:1407096:1407758 [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 cnode7-018:1407096:1407758 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407099:1407783 [3] NCCL INFO comm 0x55558141d780 rank 3 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407099:1407783 [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 cnode7-018:1407099:1407783 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407100:1407759 [4] NCCL INFO comm 0x55557c01bc00 rank 4 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407100:1407759 [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 cnode7-018:1407100:1407759 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739574:1740207 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739574:1740207 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-019:1739574:1740207 [0] NCCL INFO comm 0x555564db6160 rank 8 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-019:1739574:1740207 [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 cnode7-019:1739574:1740207 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739579:1740205 [5] NCCL INFO comm 0x555560c1a8b0 rank 13 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-019:1739579:1740205 [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 cnode7-019:1739579:1740205 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739580:1740209 [6] NCCL INFO comm 0x55557c420110 rank 14 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-019:1739580:1740209 [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 cnode7-019:1739580:1740209 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739575:1740208 [1] NCCL INFO comm 0x555561a25020 rank 9 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-019:1739575:1740208 [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 cnode7-019:1739575:1740208 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739576:1740203 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739576:1740203 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-019:1739576:1740203 [2] NCCL INFO comm 0x55556e619a10 rank 10 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-019:1739576:1740203 [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 cnode7-019:1739576:1740203 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739581:1740202 [7] NCCL INFO comm 0x555580e1fe30 rank 15 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-019:1739581:1740202 [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 cnode7-019:1739581:1740202 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739577:1740206 [3] NCCL INFO comm 0x55558161bf60 rank 11 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-019:1739577:1740206 [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 cnode7-019:1739577:1740206 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739578:1740204 [4] NCCL INFO comm 0x55557261e970 rank 12 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-019:1739578:1740204 [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 cnode7-019:1739578:1740204 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407098:1407767 [2] NCCL INFO comm 0x555561a25360 rank 2 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407098:1407767 [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 cnode7-018:1407098:1407767 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407102:1407766 [6] NCCL INFO comm 0x55557261e440 rank 6 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407102:1407766 [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 cnode7-018:1407102:1407766 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407103:1407771 [7] NCCL INFO comm 0x55557261d7b0 rank 7 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407103:1407771 [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 cnode7-018:1407103:1407771 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407101:1407851 [5] NCCL INFO comm 0x55557c41c200 rank 5 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407101:1407851 [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 cnode7-018:1407101:1407851 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739575:1740208 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 10/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 08/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 12/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 08/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 14/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Connected all rings cnode7-019:1739577:1740206 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Connected all rings cnode7-019:1739576:1740203 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Connected all rings cnode7-018:1407099:1407783 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Connected all rings cnode7-018:1407098:1407767 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Connected all rings cnode7-018:1407101:1407851 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Connected all rings cnode7-019:1739578:1740204 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Connected all rings cnode7-019:1739575:1740208 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1740208 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Connected all rings cnode7-018:1407096:1407758 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1407758 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Connected all rings cnode7-018:1407102:1407766 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1407766 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1740206 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Connected all rings cnode7-018:1407097:1407760 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1407760 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Connected all rings cnode7-018:1407103:1407771 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Connected all rings cnode7-019:1739574:1740207 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1740207 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1740203 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Connected all rings cnode7-019:1739581:1740202 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1740202 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1407783 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Connected all rings cnode7-019:1739580:1740209 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1740209 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Connected all rings cnode7-018:1407100:1407759 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1407759 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Connected all rings cnode7-019:1739579:1740205 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1740205 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1407851 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1740204 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1407771 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407098:1407767 [2] NCCL INFO Connected all trees cnode7-019:1739576:1740203 [2] NCCL INFO Connected all trees cnode7-019:1739576:1740203 [2] NCCL INFO NVLS comm 0x55556e619a10 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739577:1740206 [3] NCCL INFO Connected all trees cnode7-018:1407099:1407783 [3] NCCL INFO Connected all trees cnode7-019:1739577:1740206 [3] NCCL INFO NVLS comm 0x55558161bf60 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739578:1740204 [4] NCCL INFO Connected all trees cnode7-018:1407100:1407759 [4] NCCL INFO Connected all trees cnode7-018:1407101:1407851 [5] NCCL INFO Connected all trees cnode7-018:1407096:1407758 [0] NCCL INFO Connected all trees cnode7-018:1407097:1407760 [1] NCCL INFO Connected all trees cnode7-019:1739579:1740205 [5] NCCL INFO Connected all trees cnode7-019:1739574:1740207 [0] NCCL INFO Connected all trees cnode7-019:1739575:1740208 [1] NCCL INFO Connected all trees cnode7-018:1407103:1407771 [7] NCCL INFO Connected all trees cnode7-019:1739580:1740209 [6] NCCL INFO Connected all trees cnode7-019:1739581:1740202 [7] NCCL INFO Connected all trees cnode7-018:1407102:1407766 [6] NCCL INFO Connected all trees cnode7-019:1739581:1740202 [7] NCCL INFO NVLS comm 0x555580e1fe30 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407099:1407783 [3] NCCL INFO NVLS comm 0x55558141d780 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739580:1740209 [6] NCCL INFO NVLS comm 0x55557c420110 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407100:1407759 [4] NCCL INFO NVLS comm 0x55557c01bc00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739579:1740205 [5] NCCL INFO NVLS comm 0x555560c1a8b0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407098:1407767 [2] NCCL INFO NVLS comm 0x555561a25360 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407101:1407851 [5] NCCL INFO NVLS comm 0x55557c41c200 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739578:1740204 [4] NCCL INFO NVLS comm 0x55557261e970 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739575:1740208 [1] NCCL INFO NVLS comm 0x555561a25020 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407096:1407758 [0] NCCL INFO NVLS comm 0x5555637cce40 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407102:1407766 [6] NCCL INFO NVLS comm 0x55557261e440 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407097:1407760 [1] NCCL INFO NVLS comm 0x55556c41b400 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407103:1407771 [7] NCCL INFO NVLS comm 0x55557261d7b0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739574:1740207 [0] NCCL INFO NVLS comm 0x555564db6160 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739576:1740203 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1407783 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1407759 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1407767 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1407851 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1740208 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1407758 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1407766 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1740206 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1407771 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1740207 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1740203 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1740202 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1740209 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1740205 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1740204 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407097:1407760 [1] NCCL INFO Connected NVLS tree cnode7-018:1407097:1407760 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407097:1407760 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407103:1407771 [7] NCCL INFO Connected NVLS tree cnode7-018:1407103:1407771 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407103:1407771 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739574:1740207 [0] NCCL INFO Connected NVLS tree cnode7-019:1739574:1740207 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739574:1740207 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739575:1740208 [1] NCCL INFO Connected NVLS tree cnode7-019:1739575:1740208 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739575:1740208 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739579:1740205 [5] NCCL INFO Connected NVLS tree cnode7-019:1739579:1740205 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739579:1740205 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739576:1740203 [2] NCCL INFO Connected NVLS tree cnode7-019:1739576:1740203 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739576:1740203 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407098:1407767 [2] NCCL INFO Connected NVLS tree cnode7-018:1407098:1407767 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407098:1407767 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739580:1740209 [6] NCCL INFO Connected NVLS tree cnode7-019:1739580:1740209 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739580:1740209 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739577:1740206 [3] NCCL INFO Connected NVLS tree cnode7-019:1739577:1740206 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739577:1740206 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407101:1407851 [5] NCCL INFO Connected NVLS tree cnode7-018:1407101:1407851 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407101:1407851 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739581:1740202 [7] NCCL INFO Connected NVLS tree cnode7-019:1739581:1740202 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739581:1740202 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407099:1407783 [3] NCCL INFO Connected NVLS tree cnode7-018:1407099:1407783 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407099:1407783 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739578:1740204 [4] NCCL INFO Connected NVLS tree cnode7-019:1739578:1740204 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739578:1740204 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407100:1407759 [4] NCCL INFO Connected NVLS tree cnode7-018:1407100:1407759 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407100:1407759 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739579:1740205 [5] NCCL INFO comm 0x555560c1a8b0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407096:1407758 [0] NCCL INFO Connected NVLS tree cnode7-018:1407096:1407758 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407096:1407758 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739578:1740204 [4] NCCL INFO comm 0x55557261e970 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739574:1740207 [0] NCCL INFO comm 0x555564db6160 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739575:1740208 [1] NCCL INFO comm 0x555561a25020 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739580:1740209 [6] NCCL INFO comm 0x55557c420110 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739581:1740202 [7] NCCL INFO comm 0x555580e1fe30 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739576:1740203 [2] NCCL INFO comm 0x55556e619a10 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-019:1739577:1740206 [3] NCCL INFO comm 0x55558161bf60 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407102:1407766 [6] NCCL INFO Connected NVLS tree cnode7-018:1407102:1407766 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407102:1407766 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407098:1407767 [2] NCCL INFO comm 0x555561a25360 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407102:1407766 [6] NCCL INFO comm 0x55557261e440 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407096:1407758 [0] NCCL INFO comm 0x5555637cce40 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407097:1407760 [1] NCCL INFO comm 0x55556c41b400 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407100:1407759 [4] NCCL INFO comm 0x55557c01bc00 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407103:1407771 [7] NCCL INFO comm 0x55557261d7b0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407101:1407851 [5] NCCL INFO comm 0x55557c41c200 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0xa4a394c9439e5d9f - Init COMPLETE cnode7-018:1407099:1407783 [3] NCCL INFO comm 0x55558141d780 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0xa4a394c9439e5d9f - Init COMPLETE [2024-12-10 17:29:39,630] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Flops Profiler Enabled: False [2024-12-10 17:29:39,637] [INFO] [logging.py:129:log_dist] [Rank 0] Using DeepSpeed Optimizer param name adam as basic optimizer [2024-12-10 17:29:39,637] [INFO] [logging.py:129:log_dist] [Rank 0] Removing param_group that has no 'params' in the basic Optimizer [2024-12-10 17:29:39,675] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Basic Optimizer = FusedAdam [2024-12-10 17:29:39,675] [INFO] [utils.py:59:is_zero_supported_optimizer] Checking ZeRO support for optimizer=FusedAdam type= [2024-12-10 17:29:39,675] [INFO] [logging.py:129:log_dist] [Rank 0] Creating torch.bfloat16 ZeRO stage 2 optimizer [2024-12-10 17:29:39,675] [INFO] [stage_1_and_2.py:149:__init__] Reduce bucket size 500000000 [2024-12-10 17:29:39,675] [INFO] [stage_1_and_2.py:150:__init__] Allgather bucket size 500000000 [2024-12-10 17:29:39,675] [INFO] [stage_1_and_2.py:151:__init__] CPU Offload: False [2024-12-10 17:29:39,675] [INFO] [stage_1_and_2.py:152:__init__] Round robin gradient partitioning: False [2024-12-10 17:29:41,257] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,258] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,258] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,270] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,270] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,603] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,603] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,603] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:41,603] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,285] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,620] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,301] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,643] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,506] [INFO] [utils.py:781:see_memory_usage] Before initializing optimizer states [2024-12-10 17:29:41,525] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.45 GB CA 1.45 GB Max_CA 1 GB [2024-12-10 17:29:41,525] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 56.08 GB, percent = 2.8% [2024-12-10 17:29:41,970] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:42,005] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:41,737] [INFO] [utils.py:781:see_memory_usage] After initializing optimizer states [2024-12-10 17:29:41,738] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.52 GB CA 1.6 GB Max_CA 2 GB [2024-12-10 17:29:41,738] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 58.08 GB, percent = 2.9% [2024-12-10 17:29:41,738] [INFO] [stage_1_and_2.py:544:__init__] optimizer state initialized [2024-12-10 17:29:41,923] [INFO] [utils.py:781:see_memory_usage] After initializing ZeRO optimizer [2024-12-10 17:29:41,923] [INFO] [utils.py:782:see_memory_usage] MA 1.37 GB Max_MA 1.37 GB CA 1.6 GB Max_CA 2 GB [2024-12-10 17:29:41,923] [INFO] [utils.py:789:see_memory_usage] CPU Virtual Memory: used = 58.9 GB, percent = 2.9% [2024-12-10 17:29:41,925] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed Final Optimizer = DeepSpeedZeroOptimizer [2024-12-10 17:29:41,925] [INFO] [logging.py:129:log_dist] [Rank 0] DeepSpeed using client callable to create LR scheduler [2024-12-10 17:29:41,925] [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]) [2024-12-10 17:29:41,925] [INFO] [logging.py:129:log_dist] [Rank 0] step=0, skipped=0, lr=[np.float64(1.6666666666666667e-09)], mom=[[0.9, 0.98]] [2024-12-10 17:29:41,926] [INFO] [config.py:999:print] DeepSpeedEngine configuration: [2024-12-10 17:29:41,926] [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 } [2024-12-10 17:29:41,926] [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} [2024-12-10 17:29:41,926] [INFO] [config.py:1003:print] amp_enabled .................. False [2024-12-10 17:29:41,926] [INFO] [config.py:1003:print] amp_params ................... False [2024-12-10 17:29:41,927] [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 } [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] bfloat16_enabled ............. True [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] bfloat16_immediate_grad_update False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] checkpoint_parallel_write_pipeline False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] checkpoint_tag_validation_enabled True [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] checkpoint_tag_validation_fail False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] comms_config ................. [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] communication_data_type ...... None [2024-12-10 17:29:41,927] [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}} [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] curriculum_enabled_legacy .... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] curriculum_params_legacy ..... False [2024-12-10 17:29:41,927] [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}}}} [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] data_efficiency_enabled ...... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] dataloader_drop_last ......... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] disable_allgather ............ False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] dump_state ................... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] dynamic_loss_scale_args ...... None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_enabled ........... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_gas_boundary_resolution 1 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_layer_name ........ bert.encoder.layer [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_layer_num ......... 0 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_max_iter .......... 100 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_stability ......... 1e-06 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_tol ............... 0.01 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] eigenvalue_verbose ........... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] elasticity_enabled ........... False [2024-12-10 17:29:41,927] [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 } [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] fp16_auto_cast ............... None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] fp16_enabled ................. False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] fp16_master_weights_and_gradients False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] global_rank .................. 0 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] grad_accum_dtype ............. None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] gradient_accumulation_steps .. 1 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] gradient_clipping ............ 5.0 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] gradient_predivide_factor .... 1.0 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] graph_harvesting ............. False [2024-12-10 17:29:41,927] [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 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] initial_dynamic_scale ........ 1 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] load_universal_checkpoint .... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] loss_scale ................... 1.0 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] memory_breakdown ............. False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] mics_hierarchial_params_gather False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] mics_shard_size .............. -1 [2024-12-10 17:29:41,927] [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') [2024-12-10 17:29:41,927] [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 } [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] optimizer_legacy_fusion ...... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] optimizer_name ............... adam [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] optimizer_params ............. {'lr': 0.0002, 'betas': [0.9, 0.98], 'eps': 1e-06, 'weight_decay': 0.0} [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] pipeline ..................... {'stages': 'auto', 'partition': 'best', 'seed_layers': False, 'activation_checkpoint_interval': 0, 'pipe_partitioned': True, 'grad_partitioned': True} [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] pld_enabled .................. False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] pld_params ................... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] prescale_gradients ........... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] scheduler_name ............... None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] scheduler_params ............. None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] seq_parallel_communication_data_type torch.float32 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] sparse_attention ............. None [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] sparse_gradients_enabled ..... False [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] steps_per_print .............. 1000 [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] timers_config ................ enabled=True synchronized=True [2024-12-10 17:29:41,927] [INFO] [config.py:1003:print] train_batch_size ............. 16 [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] train_micro_batch_size_per_gpu 1 [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] use_data_before_expert_parallel_ False [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] use_node_local_storage ....... False [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] wall_clock_breakdown ......... False [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] weight_quantization_config ... None [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] world_size ................... 16 [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] zero_allow_untested_optimizer False [2024-12-10 17:29:41,928] [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 [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] zero_enabled ................. True [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] zero_force_ds_cpu_optimizer .. True [2024-12-10 17:29:41,928] [INFO] [config.py:1003:print] zero_optimization_stage ...... 2 [2024-12-10 17:29:41,928] [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 } [cnode7-018:0/16] 2024-12-10 17:29:41,933 (deepspeed_trainer:75) INFO: Resume training from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40 [2024-12-10 17:29:41,936] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... /mnt/home/williamchen/espnet/tools/miniconda/envs/espnet/lib/python3.10/site-packages/deepspeed/runtime/checkpoint_engine/torch_checkpoint_engine.py:28: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. partition = torch.load(path, map_location=map_location) [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,779] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,848] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,849] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,853] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,854] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,854] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,854] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,855] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,858] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,596] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:42,659] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,659] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,664] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,665] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,665] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,665] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,666] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:42,673] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt... [2024-12-10 17:29:43,533] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,533] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,533] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,533] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,534] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,534] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,536] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,537] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,351] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,352] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,352] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,352] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,353] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,354] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,354] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,355] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/mp_rank_00_model_states.pt. [2024-12-10 17:29:43,757] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,761] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,793] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,793] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,794] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,806] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,809] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,831] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,649] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,650] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-10 17:29:43,650] [INFO] [torch_checkpoint_engine.py:27:load] [Torch] Loading checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-10 17:29:44,122] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,122] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 15 [2024-12-10 17:29:44,284] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 15 [2024-12-10 17:29:44,291] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,292] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 10 [2024-12-10 17:29:44,351] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 10 [2024-12-10 17:29:44,359] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,360] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 11 [2024-12-10 17:29:44,055] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,056] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 7 [2024-12-10 17:29:44,423] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 11 [2024-12-10 17:29:44,114] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,114] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 6 [2024-12-10 17:29:44,123] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 7 [2024-12-10 17:29:44,455] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,455] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 8 [2024-12-10 17:29:44,466] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,466] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 12 [2024-12-10 17:29:44,474] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,475] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 14 [2024-12-10 17:29:44,170] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 6 [2024-12-10 17:29:44,518] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 8 [2024-12-10 17:29:44,529] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 12 [2024-12-10 17:29:44,234] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,235] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 4 [2024-12-10 17:29:44,566] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,567] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 13 [2024-12-10 17:29:44,593] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,593] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 9 [2024-12-10 17:29:44,596] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 14 [2024-12-10 17:29:44,288] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,288] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 1 [2024-12-10 17:29:44,290] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 4 [2024-12-10 17:29:44,622] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 13 [2024-12-10 17:29:44,297] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,297] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 5 [2024-12-10 17:29:44,649] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 9 [2024-12-10 17:29:44,376] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,377] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 2 [2024-12-10 17:29:44,377] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 5 [2024-12-10 17:29:44,377] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,378] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 3 [2024-12-10 17:29:44,378] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 1 [2024-12-10 17:29:44,411] [INFO] [torch_checkpoint_engine.py:29:load] [Torch] Loaded checkpoint from exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_40/40/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-10 17:29:44,411] [INFO] [engine.py:3076:_get_all_zero_checkpoint_state_dicts] successfully read 16 ZeRO state_dicts for rank 0 [2024-12-10 17:29:44,432] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 3 [2024-12-10 17:29:44,455] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 2 [2024-12-10 17:29:44,464] [INFO] [engine.py:3026:_load_zero_checkpoint] loading 16 zero partition checkpoints for rank 0 [cnode7-018:0/16] 2024-12-10 17:29:44,488 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-018:0/16] 2024-12-10 17:30:10,957 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 17:30:27,940 (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: ) [cnode7-018:0/16] 2024-12-10 17:30:27,940 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 17:30:27,942 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 17:30:49,461] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:49,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:50,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:50,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:50,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:50,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:50,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:51,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:51,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:51,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:51,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:51,932] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:52,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:52,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:52,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:30:52,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:36,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:36,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:37,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:37,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:38,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:38,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:39,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:39,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:39,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:39,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:40,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:40,833] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:41,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:41,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:42,873] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:31:51,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:22,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:24,361] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:25,145] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:26,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:27,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:27,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:27,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:27,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:28,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:28,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:29,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:29,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:30,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:32,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:37,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:32:39,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:08,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:10,639] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:11,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:14,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:15,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:15,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:16,169] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:17,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:16,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:16,877] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:17,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:17,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:17,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:20,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:28,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:33:30,953] [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): /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): /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): /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): /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): cnode7-018:1407102:1416445 [6] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407102:1416445 [6] NCCL INFO Using network IB cnode7-018:1407096:1416441 [0] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407096:1416441 [0] NCCL INFO Using network IB cnode7-018:1407097:1416439 [1] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407097:1416439 [1] NCCL INFO Using network IB cnode7-018:1407103:1416440 [7] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407103:1416440 [7] NCCL INFO Using network IB cnode7-019:1739576:1748607 [2] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739576:1748607 [2] NCCL INFO Using network IB cnode7-019:1739581:1748608 [7] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739581:1748608 [7] NCCL INFO Using network IB cnode7-018:1407100:1416442 [4] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407100:1416442 [4] NCCL INFO Using network IB cnode7-018:1407099:1416443 [3] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407099:1416443 [3] NCCL INFO Using network IB cnode7-019:1739578:1748605 [4] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739578:1748605 [4] NCCL INFO Using network IB cnode7-019:1739579:1748609 [5] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739579:1748609 [5] NCCL INFO Using network IB cnode7-018:1407101:1416444 [5] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407101:1416444 [5] NCCL INFO Using network IB cnode7-019:1739575:1748606 [1] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739575:1748606 [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): /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): cnode7-018:1407098:1416520 [2] NCCL INFO Using non-device net plugin version 0 cnode7-018:1407098:1416520 [2] NCCL INFO Using network IB cnode7-019:1739580:1748768 [6] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739580:1748768 [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): cnode7-019:1739577:1749185 [3] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739577:1749185 [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): cnode7-018:1407096:1416441 [0] NCCL INFO comm 0x5555b6f0b740 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739577:1749185 [3] NCCL INFO comm 0x5555d4f25c20 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407097:1416439 [1] NCCL INFO comm 0x5555c0f0d730 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407103:1416440 [7] NCCL INFO comm 0x5555c0efc8e0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739574:1749329 [0] NCCL INFO Using non-device net plugin version 0 cnode7-019:1739574:1749329 [0] NCCL INFO Using network IB cnode7-019:1739574:1749329 [0] NCCL INFO comm 0x5555b6f4afd0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739576:1748607 [2] NCCL INFO comm 0x5555c0f343b0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739581:1748608 [7] NCCL INFO comm 0x5555c0eeb670 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407100:1416442 [4] NCCL INFO comm 0x5555c0f3b0b0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407099:1416443 [3] NCCL INFO comm 0x5555c0f187e0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407098:1416520 [2] NCCL INFO comm 0x5555c0f2f350 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739578:1748605 [4] NCCL INFO comm 0x5555c0f24ef0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739579:1748609 [5] NCCL INFO comm 0x5555c0f35550 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407101:1416444 [5] NCCL INFO comm 0x5555d4f1fac0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739575:1748606 [1] NCCL INFO comm 0x5555be779450 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x25f4325fe90625b5 - Init START cnode7-018:1407102:1416445 [6] NCCL INFO comm 0x5555c0eea530 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739580:1748768 [6] NCCL INFO comm 0x5555c0f382e0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x25f4325fe90625b5 - Init START cnode7-019:1739574:1749329 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739574:1749329 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-019:1739577:1749185 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739577:1749185 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-019:1739580:1748768 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739580:1748768 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-019:1739575:1748606 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739575:1748606 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-018:1407102:1416445 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS multicast support is available on dev 6 cnode7-018:1407103:1416440 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-019:1739576:1748607 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-019:1739576:1748607 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-019:1739581:1748608 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739581:1748608 [7] NCCL INFO NVLS multicast support is available on dev 7 cnode7-019:1739578:1748605 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739578:1748605 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-019:1739579:1748609 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-019:1739579:1748609 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-018:1407101:1416444 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS multicast support is available on dev 5 cnode7-018:1407096:1416441 [0] NCCL INFO Setting affinity for GPU 0 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407096:1416441 [0] NCCL INFO NVLS multicast support is available on dev 0 cnode7-018:1407097:1416439 [1] NCCL INFO Setting affinity for GPU 1 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407097:1416439 [1] NCCL INFO NVLS multicast support is available on dev 1 cnode7-018:1407100:1416442 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffffff00,00000000,0000ffff,ffffffff,ff000000,00000000 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS multicast support is available on dev 4 cnode7-018:1407100:1416442 [4] NCCL INFO comm 0x5555c0f3b0b0 rank 4 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407100:1416442 [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 cnode7-018:1407100:1416442 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407101:1416444 [5] NCCL INFO comm 0x5555d4f1fac0 rank 5 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407101:1416444 [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 cnode7-018:1407101:1416444 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407103:1416440 [7] NCCL INFO comm 0x5555c0efc8e0 rank 7 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407103:1416440 [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 cnode7-018:1407103:1416440 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407102:1416445 [6] NCCL INFO comm 0x5555c0eea530 rank 6 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407102:1416445 [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 cnode7-018:1407102:1416445 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407099:1416443 [3] NCCL INFO Setting affinity for GPU 3 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407099:1416443 [3] NCCL INFO NVLS multicast support is available on dev 3 cnode7-018:1407099:1416443 [3] NCCL INFO comm 0x5555c0f187e0 rank 3 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407099:1416443 [3] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407099:1416443 [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 cnode7-018:1407099:1416443 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739579:1748609 [5] NCCL INFO comm 0x5555c0f35550 rank 13 nRanks 16 nNodes 2 localRanks 8 localRank 5 MNNVL 0 cnode7-019:1739579:1748609 [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 cnode7-019:1739579:1748609 [5] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739578:1748605 [4] NCCL INFO comm 0x5555c0f24ef0 rank 12 nRanks 16 nNodes 2 localRanks 8 localRank 4 MNNVL 0 cnode7-019:1739578:1748605 [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 cnode7-019:1739578:1748605 [4] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739577:1749185 [3] NCCL INFO comm 0x5555d4f25c20 rank 11 nRanks 16 nNodes 2 localRanks 8 localRank 3 MNNVL 0 cnode7-019:1739577:1749185 [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 cnode7-019:1739577:1749185 [3] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739575:1748606 [1] NCCL INFO comm 0x5555be779450 rank 9 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-019:1739575:1748606 [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 cnode7-019:1739575:1748606 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739576:1748607 [2] NCCL INFO comm 0x5555c0f343b0 rank 10 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-019:1739576:1748607 [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 cnode7-019:1739576:1748607 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739574:1749329 [0] NCCL INFO comm 0x5555b6f4afd0 rank 8 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-019:1739574:1749329 [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 cnode7-019:1739574:1749329 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739580:1748768 [6] NCCL INFO comm 0x5555c0f382e0 rank 14 nRanks 16 nNodes 2 localRanks 8 localRank 6 MNNVL 0 cnode7-019:1739580:1748768 [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 cnode7-019:1739580:1748768 [6] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407097:1416439 [1] NCCL INFO comm 0x5555c0f0d730 rank 1 nRanks 16 nNodes 2 localRanks 8 localRank 1 MNNVL 0 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407097:1416439 [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 cnode7-018:1407097:1416439 [1] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407098:1416520 [2] NCCL INFO Setting affinity for GPU 2 to ff,ffffffff,ffff0000,00000000,00ffffff,ffffffff cnode7-018:1407098:1416520 [2] NCCL INFO NVLS multicast support is available on dev 2 cnode7-018:1407098:1416520 [2] NCCL INFO comm 0x5555c0f2f350 rank 2 nRanks 16 nNodes 2 localRanks 8 localRank 2 MNNVL 0 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407098:1416520 [2] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407098:1416520 [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 cnode7-018:1407098:1416520 [2] NCCL INFO P2P Chunksize set to 131072 cnode7-018:1407096:1416441 [0] NCCL INFO comm 0x5555b6f0b740 rank 0 nRanks 16 nNodes 2 localRanks 8 localRank 0 MNNVL 0 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 0: 0 8 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 1: 1 9 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 2: 2 10 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 3: 3 11 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 4: 4 12 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 5: 5 13 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 6: 6 14 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS Head 7: 7 15 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 00/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 01/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 03/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 04/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 05/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 06/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 07/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 08/16 : 0 7 6 5 4 3 2 1 9 10 11 12 13 14 15 8 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 09/16 : 0 8 15 14 13 12 11 10 9 1 2 3 4 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 11 12 13 14 15 8 9 10 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 11/16 : 0 1 2 10 9 8 15 14 13 12 11 3 4 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 12/16 : 0 7 6 5 13 14 15 8 9 10 11 12 4 3 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 13/16 : 0 1 2 3 4 12 11 10 9 8 15 14 13 5 6 7 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 14/16 : 0 7 15 8 9 10 11 12 13 14 6 5 4 3 2 1 cnode7-018:1407096:1416441 [0] NCCL INFO Channel 15/16 : 0 1 2 3 4 5 6 14 13 12 11 10 9 8 15 7 cnode7-018:1407096:1416441 [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 cnode7-018:1407096:1416441 [0] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739581:1748608 [7] NCCL INFO comm 0x5555c0eeb670 rank 15 nRanks 16 nNodes 2 localRanks 8 localRank 7 MNNVL 0 cnode7-019:1739581:1748608 [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 cnode7-019:1739581:1748608 [7] NCCL INFO P2P Chunksize set to 131072 cnode7-019:1739575:1748606 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 10/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 08/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 12/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 14/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 08/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Connected all rings cnode7-018:1407096:1416441 [0] NCCL INFO Connected all rings cnode7-018:1407096:1416441 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Connected all rings cnode7-019:1739577:1749185 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Connected all rings cnode7-018:1407097:1416439 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Connected all rings cnode7-018:1407103:1416440 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-018:1407103:1416440 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Connected all rings cnode7-019:1739574:1749329 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/CUMEM cnode7-019:1739574:1749329 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Connected all rings cnode7-019:1739576:1748607 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739576:1748607 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Connected all rings cnode7-019:1739581:1748608 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Connected all rings cnode7-018:1407100:1416442 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407100:1416442 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Connected all rings cnode7-018:1407099:1416443 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Connected all rings cnode7-018:1407098:1416520 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-018:1407098:1416520 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Connected all rings cnode7-019:1739579:1748609 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739579:1748609 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Connected all rings cnode7-019:1739578:1748605 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-019:1739578:1748605 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Connected all rings cnode7-018:1407101:1416444 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-018:1407101:1416444 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Connected all rings cnode7-019:1739575:1748606 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-019:1739575:1748606 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Connected all rings cnode7-018:1407102:1416445 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-018:1407102:1416445 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-019:1739580:1748768 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/CUMEM cnode7-018:1407096:1416441 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-019:1739577:1749185 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-018:1407097:1416439 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-019:1739581:1748608 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/CUMEM cnode7-018:1407099:1416443 [3] NCCL INFO Connected all trees cnode7-019:1739577:1749185 [3] NCCL INFO Connected all trees cnode7-018:1407099:1416443 [3] NCCL INFO NVLS comm 0x5555c0f187e0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407098:1416520 [2] NCCL INFO Connected all trees cnode7-019:1739576:1748607 [2] NCCL INFO Connected all trees cnode7-019:1739577:1749185 [3] NCCL INFO NVLS comm 0x5555d4f25c20 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407101:1416444 [5] NCCL INFO Connected all trees cnode7-018:1407100:1416442 [4] NCCL INFO Connected all trees cnode7-019:1739578:1748605 [4] NCCL INFO Connected all trees cnode7-019:1739579:1748609 [5] NCCL INFO Connected all trees cnode7-019:1739581:1748608 [7] NCCL INFO Connected all trees cnode7-018:1407102:1416445 [6] NCCL INFO Connected all trees cnode7-019:1739580:1748768 [6] NCCL INFO Connected all trees cnode7-018:1407096:1416441 [0] NCCL INFO Connected all trees cnode7-018:1407103:1416440 [7] NCCL INFO Connected all trees cnode7-018:1407097:1416439 [1] NCCL INFO Connected all trees cnode7-019:1739574:1749329 [0] NCCL INFO Connected all trees cnode7-019:1739576:1748607 [2] NCCL INFO NVLS comm 0x5555c0f343b0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739575:1748606 [1] NCCL INFO Connected all trees cnode7-018:1407098:1416520 [2] NCCL INFO NVLS comm 0x5555c0f2f350 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739579:1748609 [5] NCCL INFO NVLS comm 0x5555c0f35550 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739578:1748605 [4] NCCL INFO NVLS comm 0x5555c0f24ef0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407101:1416444 [5] NCCL INFO NVLS comm 0x5555d4f1fac0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739575:1748606 [1] NCCL INFO NVLS comm 0x5555be779450 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407102:1416445 [6] NCCL INFO NVLS comm 0x5555c0eea530 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739580:1748768 [6] NCCL INFO NVLS comm 0x5555c0f382e0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407096:1416441 [0] NCCL INFO NVLS comm 0x5555b6f0b740 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407097:1416439 [1] NCCL INFO NVLS comm 0x5555c0f0d730 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407103:1416440 [7] NCCL INFO NVLS comm 0x5555c0efc8e0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407103:1416440 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407103:1416440 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO NVLS comm 0x5555b6f4afd0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-019:1739581:1748608 [7] NCCL INFO NVLS comm 0x5555c0eeb670 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407100:1416442 [4] NCCL INFO NVLS comm 0x5555c0f3b0b0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 cnode7-018:1407100:1416442 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407100:1416442 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407099:1416443 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739579:1748609 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-019:1739578:1748605 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/IB/4/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-018:1407101:1416444 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/IB/6/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739575:1748606 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407102:1416445 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-019:1739580:1748768 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/IB/7/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-018:1407096:1416441 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-019:1739577:1749185 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/IB/3/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-018:1407097:1416439 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/IB/1/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739574:1749329 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/IB/0/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739576:1748607 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/IB/2/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-019:1739581:1748608 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/IB/8/GDRDMA cnode7-018:1407098:1416520 [2] NCCL INFO Connected NVLS tree cnode7-018:1407098:1416520 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407098:1416520 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739574:1749329 [0] NCCL INFO Connected NVLS tree cnode7-019:1739574:1749329 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739574:1749329 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739579:1748609 [5] NCCL INFO Connected NVLS tree cnode7-019:1739579:1748609 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739579:1748609 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407101:1416444 [5] NCCL INFO Connected NVLS tree cnode7-018:1407101:1416444 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407101:1416444 [5] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739576:1748607 [2] NCCL INFO Connected NVLS tree cnode7-019:1739576:1748607 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739576:1748607 [2] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739580:1748768 [6] NCCL INFO Connected NVLS tree cnode7-019:1739580:1748768 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739580:1748768 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739577:1749185 [3] NCCL INFO Connected NVLS tree cnode7-019:1739577:1749185 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739577:1749185 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739575:1748606 [1] NCCL INFO Connected NVLS tree cnode7-019:1739575:1748606 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739575:1748606 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407096:1416441 [0] NCCL INFO Connected NVLS tree cnode7-018:1407096:1416441 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407096:1416441 [0] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407102:1416445 [6] NCCL INFO Connected NVLS tree cnode7-018:1407102:1416445 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407102:1416445 [6] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407097:1416439 [1] NCCL INFO Connected NVLS tree cnode7-018:1407097:1416439 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407097:1416439 [1] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739581:1748608 [7] NCCL INFO Connected NVLS tree cnode7-019:1739581:1748608 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739581:1748608 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739578:1748605 [4] NCCL INFO Connected NVLS tree cnode7-019:1739578:1748605 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-019:1739578:1748605 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407103:1416440 [7] NCCL INFO Connected NVLS tree cnode7-018:1407103:1416440 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407103:1416440 [7] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407100:1416442 [4] NCCL INFO Connected NVLS tree cnode7-018:1407100:1416442 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407100:1416442 [4] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-018:1407099:1416443 [3] NCCL INFO Connected NVLS tree cnode7-018:1407099:1416443 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 cnode7-018:1407099:1416443 [3] NCCL INFO 16 coll channels, 0 collnet channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer cnode7-019:1739577:1749185 [3] NCCL INFO comm 0x5555d4f25c20 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407097:1416439 [1] NCCL INFO comm 0x5555c0f0d730 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407096:1416441 [0] NCCL INFO comm 0x5555b6f0b740 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407103:1416440 [7] NCCL INFO comm 0x5555c0efc8e0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407102:1416445 [6] NCCL INFO comm 0x5555c0eea530 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407099:1416443 [3] NCCL INFO comm 0x5555c0f187e0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 61000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407101:1416444 [5] NCCL INFO comm 0x5555d4f1fac0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407098:1416520 [2] NCCL INFO comm 0x5555c0f2f350 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-018:1407100:1416442 [4] NCCL INFO comm 0x5555c0f3b0b0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739574:1749329 [0] NCCL INFO comm 0x5555b6f4afd0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 1b000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739579:1748609 [5] NCCL INFO comm 0x5555c0f35550 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId c3000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739578:1748605 [4] NCCL INFO comm 0x5555c0f24ef0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 9d000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739575:1748606 [1] NCCL INFO comm 0x5555be779450 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 43000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739581:1748608 [7] NCCL INFO comm 0x5555c0eeb670 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId df000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739576:1748607 [2] NCCL INFO comm 0x5555c0f343b0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 52000 commId 0x25f4325fe90625b5 - Init COMPLETE cnode7-019:1739580:1748768 [6] NCCL INFO comm 0x5555c0f382e0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId d1000 commId 0x25f4325fe90625b5 - Init COMPLETE [cnode7-018:0/16] 2024-12-10 17:34:42,263 (deepspeed_trainer:228) INFO: 41epoch:train:1-100batch: iter_time=2.407, loss_ctc=73.606, loss_att=61.032, acc=0.718, loss=64.812, grad_norm=5.811, loss_scale=1.000, learning_rate=6.324e-05, step_time=0.583 [cnode7-018:0/16] 2024-12-10 17:35:18,964 (deepspeed_trainer:228) INFO: 41epoch:train:101-200batch: iter_time=1.126e-04, loss_ctc=71.861, loss_att=54.069, acc=0.724, loss=59.400, grad_norm=6.418, loss_scale=1.000, learning_rate=6.324e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:35:55,672 (deepspeed_trainer:228) INFO: 41epoch:train:201-300batch: iter_time=1.132e-04, loss_ctc=78.678, loss_att=61.478, acc=0.709, loss=66.650, grad_norm=7.835, loss_scale=1.000, learning_rate=6.323e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:36:32,272 (deepspeed_trainer:228) INFO: 41epoch:train:301-400batch: iter_time=1.152e-04, loss_ctc=66.671, loss_att=51.922, acc=0.719, loss=56.339, grad_norm=5.917, loss_scale=1.000, learning_rate=6.323e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 17:37:09,007 (deepspeed_trainer:228) INFO: 41epoch:train:401-500batch: iter_time=1.233e-04, loss_ctc=75.424, loss_att=53.407, acc=0.714, loss=60.017, grad_norm=6.612, loss_scale=1.000, learning_rate=6.322e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:37:45,834 (deepspeed_trainer:228) INFO: 41epoch:train:501-600batch: iter_time=1.251e-04, loss_ctc=63.317, loss_att=47.015, acc=0.720, loss=51.906, grad_norm=6.346, loss_scale=1.000, learning_rate=6.322e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:38:22,703 (deepspeed_trainer:228) INFO: 41epoch:train:601-700batch: iter_time=1.204e-04, loss_ctc=67.231, loss_att=50.829, acc=0.721, loss=55.765, grad_norm=5.906, loss_scale=1.000, learning_rate=6.321e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 17:38:59,135 (deepspeed_trainer:228) INFO: 41epoch:train:701-800batch: iter_time=1.172e-04, loss_ctc=81.083, loss_att=56.418, acc=0.724, loss=63.828, grad_norm=7.276, loss_scale=1.000, learning_rate=6.321e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:39:35,728 (deepspeed_trainer:228) INFO: 41epoch:train:801-900batch: iter_time=1.183e-04, loss_ctc=78.072, loss_att=63.538, acc=0.697, loss=67.939, grad_norm=7.455, loss_scale=1.000, learning_rate=6.320e-05, step_time=0.365 [2024-12-10 17:40:12,184] [INFO] [logging.py:129:log_dist] [Rank 0] step=601000, skipped=0, lr=[np.float64(6.319286179204039e-05)], mom=[[0.9, 0.98]] [2024-12-10 17:40:12,184] [INFO] [timer.py:264:stop] epoch=0/micro_step=1000/global_step=1000, RunningAvgSamplesPerSec=44.565795448707064, CurrSamplesPerSec=43.82126904625384, MemAllocated=1.77GB, MaxMemAllocated=15.68GB [cnode7-018:0/16] 2024-12-10 17:40:12,186 (deepspeed_trainer:228) INFO: 41epoch:train:901-1000batch: iter_time=1.210e-04, loss_ctc=67.576, loss_att=53.774, acc=0.728, loss=57.937, grad_norm=5.682, loss_scale=1.000, learning_rate=6.320e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:40:48,536 (deepspeed_trainer:228) INFO: 41epoch:train:1001-1100batch: iter_time=1.120e-04, loss_ctc=66.980, loss_att=49.036, acc=0.725, loss=54.421, grad_norm=6.378, loss_scale=1.000, learning_rate=6.319e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 17:41:24,744 (deepspeed_trainer:228) INFO: 41epoch:train:1101-1200batch: iter_time=1.158e-04, loss_ctc=59.377, loss_att=43.248, acc=0.736, loss=48.089, grad_norm=5.407, loss_scale=1.000, learning_rate=6.318e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 17:42:01,002 (deepspeed_trainer:228) INFO: 41epoch:train:1201-1300batch: iter_time=1.106e-04, loss_ctc=75.137, loss_att=53.362, acc=0.730, loss=59.875, grad_norm=5.986, loss_scale=1.000, learning_rate=6.318e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 17:42:37,490 (deepspeed_trainer:228) INFO: 41epoch:train:1301-1400batch: iter_time=1.091e-04, loss_ctc=76.208, loss_att=58.701, acc=0.711, loss=63.958, grad_norm=7.089, loss_scale=1.000, learning_rate=6.317e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 17:43:14,002 (deepspeed_trainer:228) INFO: 41epoch:train:1401-1500batch: iter_time=1.097e-04, loss_ctc=77.919, loss_att=56.475, acc=0.734, loss=62.913, grad_norm=6.516, loss_scale=1.000, learning_rate=6.317e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 17:43:50,245 (deepspeed_trainer:228) INFO: 41epoch:train:1501-1600batch: iter_time=1.085e-04, loss_ctc=66.192, loss_att=49.461, acc=0.743, loss=54.502, grad_norm=5.447, loss_scale=1.000, learning_rate=6.316e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 17:44:26,318 (deepspeed_trainer:228) INFO: 41epoch:train:1601-1700batch: iter_time=1.068e-04, loss_ctc=67.398, loss_att=48.556, acc=0.728, loss=54.180, grad_norm=6.268, loss_scale=1.000, learning_rate=6.316e-05, step_time=0.360 [cnode7-018:0/16] 2024-12-10 17:45:02,724 (deepspeed_trainer:228) INFO: 41epoch:train:1701-1800batch: iter_time=1.088e-04, loss_ctc=69.826, loss_att=48.669, acc=0.738, loss=55.019, grad_norm=5.855, loss_scale=1.000, learning_rate=6.315e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:45:34,167 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-018:0/16] 2024-12-10 17:45:59,699 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 17:46:15,069 (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: ) [cnode7-018:0/16] 2024-12-10 17:46:15,069 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 17:46:15,071 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 17:46:40,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:41,236] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:42,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:42,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:42,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:43,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:43,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:43,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:43,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:43,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:44,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:44,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:45,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:45,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:45,743] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:46:49,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:27,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:29,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:29,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:30,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:31,516] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:32,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:32,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:32,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:32,389] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:32,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:33,253] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:33,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:33,788] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:34,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:34,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:47:37,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:14,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:16,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:18,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:19,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:19,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:20,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:21,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:21,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:21,664] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:21,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:21,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:22,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:22,885] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:23,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:24,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:48:25,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:01,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:03,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:05,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:07,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:08,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:09,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:09,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:10,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:10,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:10,932] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:11,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:11,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:11,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:11,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:13,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 17:49:14,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 17:49:49,897 (deepspeed_trainer:228) INFO: 41epoch:train:1801-1900batch: iter_time=2.460, loss_ctc=82.406, loss_att=62.079, acc=0.702, loss=68.169, grad_norm=7.174, loss_scale=1.000, learning_rate=6.315e-05, step_time=0.412 [2024-12-10 17:50:26,708] [INFO] [logging.py:129:log_dist] [Rank 0] step=602000, skipped=0, lr=[np.float64(6.314035429915395e-05)], mom=[[0.9, 0.98]] [2024-12-10 17:50:26,708] [INFO] [timer.py:264:stop] epoch=0/micro_step=2000/global_step=2000, RunningAvgSamplesPerSec=44.44522419138628, CurrSamplesPerSec=44.71967946737121, MemAllocated=1.77GB, MaxMemAllocated=15.69GB [cnode7-018:0/16] 2024-12-10 17:50:26,710 (deepspeed_trainer:228) INFO: 41epoch:train:1901-2000batch: iter_time=1.059e-04, loss_ctc=65.571, loss_att=53.016, acc=0.732, loss=56.775, grad_norm=5.122, loss_scale=1.000, learning_rate=6.314e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 17:51:03,832 (deepspeed_trainer:228) INFO: 41epoch:train:2001-2100batch: iter_time=1.053e-04, loss_ctc=84.679, loss_att=64.577, acc=0.718, loss=70.626, grad_norm=7.192, loss_scale=1.000, learning_rate=6.314e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 17:51:40,772 (deepspeed_trainer:228) INFO: 41epoch:train:2101-2200batch: iter_time=1.113e-04, loss_ctc=66.181, loss_att=53.065, acc=0.725, loss=56.976, grad_norm=5.840, loss_scale=1.000, learning_rate=6.313e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 17:52:17,751 (deepspeed_trainer:228) INFO: 41epoch:train:2201-2300batch: iter_time=1.140e-04, loss_ctc=71.201, loss_att=54.416, acc=0.722, loss=59.435, grad_norm=5.876, loss_scale=1.000, learning_rate=6.313e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 17:52:54,304 (deepspeed_trainer:228) INFO: 41epoch:train:2301-2400batch: iter_time=1.180e-04, loss_ctc=71.686, loss_att=48.061, acc=0.729, loss=55.149, grad_norm=6.652, loss_scale=1.000, learning_rate=6.312e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 17:53:30,747 (deepspeed_trainer:228) INFO: 41epoch:train:2401-2500batch: iter_time=1.200e-04, loss_ctc=60.679, loss_att=47.103, acc=0.726, loss=51.167, grad_norm=5.929, loss_scale=1.000, learning_rate=6.312e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:54:07,567 (deepspeed_trainer:228) INFO: 41epoch:train:2501-2600batch: iter_time=1.205e-04, loss_ctc=74.371, loss_att=56.410, acc=0.727, loss=61.801, grad_norm=5.854, loss_scale=1.000, learning_rate=6.311e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 17:54:43,994 (deepspeed_trainer:228) INFO: 41epoch:train:2601-2700batch: iter_time=1.141e-04, loss_ctc=68.287, loss_att=50.000, acc=0.721, loss=55.479, grad_norm=6.493, loss_scale=1.000, learning_rate=6.311e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:55:20,785 (deepspeed_trainer:228) INFO: 41epoch:train:2701-2800batch: iter_time=1.204e-04, loss_ctc=80.464, loss_att=64.541, acc=0.716, loss=69.297, grad_norm=8.463, loss_scale=1.000, learning_rate=6.310e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:55:57,047 (deepspeed_trainer:228) INFO: 41epoch:train:2801-2900batch: iter_time=1.216e-04, loss_ctc=63.721, loss_att=51.654, acc=0.732, loss=55.265, grad_norm=5.651, loss_scale=1.000, learning_rate=6.310e-05, step_time=0.362 [2024-12-10 17:56:33,611] [INFO] [logging.py:129:log_dist] [Rank 0] step=603000, skipped=0, lr=[np.float64(6.308797747591709e-05)], mom=[[0.9, 0.98]] [2024-12-10 17:56:33,611] [INFO] [timer.py:264:stop] epoch=0/micro_step=3000/global_step=3000, RunningAvgSamplesPerSec=44.498594426331834, CurrSamplesPerSec=45.265035455061025, MemAllocated=1.77GB, MaxMemAllocated=15.69GB [cnode7-018:0/16] 2024-12-10 17:56:33,613 (deepspeed_trainer:228) INFO: 41epoch:train:2901-3000batch: iter_time=1.171e-04, loss_ctc=66.871, loss_att=50.092, acc=0.738, loss=55.116, grad_norm=5.721, loss_scale=1.000, learning_rate=6.309e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 17:57:10,283 (deepspeed_trainer:228) INFO: 41epoch:train:3001-3100batch: iter_time=1.171e-04, loss_ctc=62.170, loss_att=46.333, acc=0.741, loss=51.062, grad_norm=5.098, loss_scale=1.000, learning_rate=6.309e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 17:57:47,054 (deepspeed_trainer:228) INFO: 41epoch:train:3101-3200batch: iter_time=1.139e-04, loss_ctc=75.054, loss_att=51.432, acc=0.737, loss=58.514, grad_norm=6.267, loss_scale=1.000, learning_rate=6.308e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 17:58:23,483 (deepspeed_trainer:228) INFO: 41epoch:train:3201-3300batch: iter_time=1.132e-04, loss_ctc=72.006, loss_att=55.707, acc=0.717, loss=60.606, grad_norm=6.062, loss_scale=1.000, learning_rate=6.307e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 17:59:01,079 (deepspeed_trainer:228) INFO: 41epoch:train:3301-3400batch: iter_time=1.130e-04, loss_ctc=80.883, loss_att=57.597, acc=0.749, loss=64.588, grad_norm=6.411, loss_scale=1.000, learning_rate=6.307e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 17:59:37,722 (deepspeed_trainer:228) INFO: 41epoch:train:3401-3500batch: iter_time=1.077e-04, loss_ctc=62.155, loss_att=49.983, acc=0.734, loss=53.628, grad_norm=5.228, loss_scale=1.000, learning_rate=6.306e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 18:00:13,486 (deepspeed_trainer:228) INFO: 41epoch:train:3501-3600batch: iter_time=1.094e-04, loss_ctc=66.719, loss_att=46.156, acc=0.737, loss=52.326, grad_norm=6.090, loss_scale=1.000, learning_rate=6.306e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 18:00:50,045 (deepspeed_trainer:228) INFO: 41epoch:train:3601-3700batch: iter_time=1.122e-04, loss_ctc=77.617, loss_att=55.727, acc=0.729, loss=62.301, grad_norm=6.763, loss_scale=1.000, learning_rate=6.305e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:01:12,410 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-018:0/16] 2024-12-10 18:01:38,371 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 18:01:55,165 (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: ) [cnode7-018:0/16] 2024-12-10 18:01:55,165 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 18:01:55,167 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 18:02:20,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:21,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:21,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:22,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:21,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:22,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:23,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:22,880] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:23,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:23,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:23,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:24,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:24,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:24,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:24,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:02:24,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:07,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:08,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:09,831] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:09,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:11,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:11,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:11,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:12,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:12,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:12,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:12,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:13,243] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:13,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:12,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:13,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:13,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:55,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:56,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:56,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:56,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:59,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:03:59,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:00,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:00,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:00,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:01,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:00,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:01,473] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:01,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:01,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:02,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:03,545] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:42,158] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:43,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:42,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:43,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:47,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:48,135] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:48,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:48,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:49,143] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:49,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:49,195] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:49,237] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:49,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:50,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:54,275] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:04:57,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 18:05:42,340 (deepspeed_trainer:228) INFO: 41epoch:train:3701-3800batch: iter_time=2.478, loss_ctc=80.159, loss_att=65.217, acc=0.710, loss=69.698, grad_norm=6.812, loss_scale=1.000, learning_rate=6.305e-05, step_time=0.445 [cnode7-018:0/16] 2024-12-10 18:06:18,979 (deepspeed_trainer:228) INFO: 41epoch:train:3801-3900batch: iter_time=1.052e-04, loss_ctc=63.451, loss_att=47.017, acc=0.735, loss=51.941, grad_norm=5.424, loss_scale=1.000, learning_rate=6.304e-05, step_time=0.366 [2024-12-10 18:06:55,892] [INFO] [logging.py:129:log_dist] [Rank 0] step=604000, skipped=0, lr=[np.float64(6.303573078125529e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:06:55,892] [INFO] [timer.py:264:stop] epoch=0/micro_step=4000/global_step=4000, RunningAvgSamplesPerSec=44.286760821397614, CurrSamplesPerSec=43.23932261648216, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:06:55,894 (deepspeed_trainer:228) INFO: 41epoch:train:3901-4000batch: iter_time=1.071e-04, loss_ctc=79.065, loss_att=61.431, acc=0.715, loss=66.698, grad_norm=7.307, loss_scale=1.000, learning_rate=6.304e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:07:32,458 (deepspeed_trainer:228) INFO: 41epoch:train:4001-4100batch: iter_time=1.152e-04, loss_ctc=63.755, loss_att=50.515, acc=0.726, loss=54.479, grad_norm=6.224, loss_scale=1.000, learning_rate=6.303e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:08:08,821 (deepspeed_trainer:228) INFO: 41epoch:train:4101-4200batch: iter_time=1.131e-04, loss_ctc=73.483, loss_att=54.288, acc=0.719, loss=60.055, grad_norm=6.025, loss_scale=1.000, learning_rate=6.303e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:08:45,150 (deepspeed_trainer:228) INFO: 41epoch:train:4201-4300batch: iter_time=1.151e-04, loss_ctc=70.164, loss_att=47.763, acc=0.724, loss=54.485, grad_norm=6.339, loss_scale=1.000, learning_rate=6.302e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 18:09:21,512 (deepspeed_trainer:228) INFO: 41epoch:train:4301-4400batch: iter_time=1.141e-04, loss_ctc=62.593, loss_att=50.420, acc=0.725, loss=54.104, grad_norm=5.583, loss_scale=1.000, learning_rate=6.302e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 18:09:58,040 (deepspeed_trainer:228) INFO: 41epoch:train:4401-4500batch: iter_time=1.116e-04, loss_ctc=73.022, loss_att=53.629, acc=0.720, loss=59.452, grad_norm=6.343, loss_scale=1.000, learning_rate=6.301e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 18:10:34,608 (deepspeed_trainer:228) INFO: 41epoch:train:4501-4600batch: iter_time=1.120e-04, loss_ctc=78.846, loss_att=55.017, acc=0.711, loss=62.164, grad_norm=8.129, loss_scale=1.000, learning_rate=6.301e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:11:11,382 (deepspeed_trainer:228) INFO: 41epoch:train:4601-4700batch: iter_time=1.100e-04, loss_ctc=76.578, loss_att=60.904, acc=0.721, loss=65.585, grad_norm=6.032, loss_scale=1.000, learning_rate=6.300e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 18:11:47,870 (deepspeed_trainer:228) INFO: 41epoch:train:4701-4800batch: iter_time=1.108e-04, loss_ctc=62.572, loss_att=49.286, acc=0.735, loss=53.279, grad_norm=6.611, loss_scale=1.000, learning_rate=6.300e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:12:24,071 (deepspeed_trainer:228) INFO: 41epoch:train:4801-4900batch: iter_time=1.125e-04, loss_ctc=62.001, loss_att=46.030, acc=0.736, loss=50.793, grad_norm=5.271, loss_scale=1.000, learning_rate=6.299e-05, step_time=0.362 [2024-12-10 18:13:00,341] [INFO] [logging.py:129:log_dist] [Rank 0] step=605000, skipped=0, lr=[np.float64(6.298361367722558e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:13:00,341] [INFO] [timer.py:264:stop] epoch=0/micro_step=5000/global_step=5000, RunningAvgSamplesPerSec=44.4072325875211, CurrSamplesPerSec=44.6085042318573, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:13:00,343 (deepspeed_trainer:228) INFO: 41epoch:train:4901-5000batch: iter_time=1.147e-04, loss_ctc=66.243, loss_att=49.707, acc=0.733, loss=54.661, grad_norm=5.912, loss_scale=1.000, learning_rate=6.299e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 18:13:36,782 (deepspeed_trainer:228) INFO: 41epoch:train:5001-5100batch: iter_time=1.092e-04, loss_ctc=75.265, loss_att=52.718, acc=0.727, loss=59.470, grad_norm=7.152, loss_scale=1.000, learning_rate=6.298e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:14:13,416 (deepspeed_trainer:228) INFO: 41epoch:train:5101-5200batch: iter_time=1.069e-04, loss_ctc=70.731, loss_att=53.832, acc=0.727, loss=58.908, grad_norm=5.723, loss_scale=1.000, learning_rate=6.298e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:14:51,261 (deepspeed_trainer:228) INFO: 41epoch:train:5201-5300batch: iter_time=1.056e-04, loss_ctc=80.584, loss_att=56.195, acc=0.741, loss=63.507, grad_norm=6.482, loss_scale=1.000, learning_rate=6.297e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 18:15:27,714 (deepspeed_trainer:228) INFO: 41epoch:train:5301-5400batch: iter_time=1.037e-04, loss_ctc=64.212, loss_att=49.651, acc=0.734, loss=54.051, grad_norm=5.264, loss_scale=1.000, learning_rate=6.297e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:16:04,262 (deepspeed_trainer:228) INFO: 41epoch:train:5401-5500batch: iter_time=1.047e-04, loss_ctc=63.280, loss_att=45.165, acc=0.733, loss=50.581, grad_norm=5.867, loss_scale=1.000, learning_rate=6.296e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:16:41,016 (deepspeed_trainer:228) INFO: 41epoch:train:5501-5600batch: iter_time=1.038e-04, loss_ctc=80.859, loss_att=58.319, acc=0.721, loss=65.080, grad_norm=6.760, loss_scale=1.000, learning_rate=6.295e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:16:53,925 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-018:0/16] 2024-12-10 18:17:20,512 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 18:17:36,671 (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: ) [cnode7-018:0/16] 2024-12-10 18:17:36,672 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 18:17:36,674 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 18:18:01,975] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:02,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:02,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:02,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:03,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:03,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:03,440] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:03,840] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:03,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,067] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:04,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:05,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:49,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:50,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:50,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:50,653] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:51,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:51,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:52,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:52,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:53,249] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:53,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:53,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:53,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:54,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:53,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:55,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:18:56,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:36,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:38,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:39,032] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:39,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:40,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:40,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:40,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:41,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:41,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:41,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:42,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:42,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:42,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:43,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:46,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:19:47,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:23,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:26,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:26,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:27,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:28,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:29,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:29,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:29,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:30,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:30,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:30,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:31,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:31,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:33,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:36,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:20:37,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 18:21:31,588 (deepspeed_trainer:228) INFO: 41epoch:train:5601-5700batch: iter_time=2.459, loss_ctc=71.103, loss_att=56.589, acc=0.721, loss=60.917, grad_norm=6.072, loss_scale=1.000, learning_rate=6.295e-05, step_time=0.446 [cnode7-018:0/16] 2024-12-10 18:22:08,734 (deepspeed_trainer:228) INFO: 41epoch:train:5701-5800batch: iter_time=1.059e-04, loss_ctc=67.959, loss_att=50.321, acc=0.731, loss=55.616, grad_norm=6.159, loss_scale=1.000, learning_rate=6.294e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 18:22:45,870 (deepspeed_trainer:228) INFO: 41epoch:train:5801-5900batch: iter_time=1.095e-04, loss_ctc=80.061, loss_att=63.388, acc=0.718, loss=68.401, grad_norm=7.537, loss_scale=1.000, learning_rate=6.294e-05, step_time=0.371 [2024-12-10 18:23:22,624] [INFO] [logging.py:129:log_dist] [Rank 0] step=606000, skipped=0, lr=[np.float64(6.293162562899316e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:23:22,625] [INFO] [timer.py:264:stop] epoch=0/micro_step=6000/global_step=6000, RunningAvgSamplesPerSec=44.24385902834934, CurrSamplesPerSec=39.6231727189181, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:23:22,627 (deepspeed_trainer:228) INFO: 41epoch:train:5901-6000batch: iter_time=1.091e-04, loss_ctc=66.125, loss_att=48.768, acc=0.734, loss=53.974, grad_norm=6.300, loss_scale=1.000, learning_rate=6.293e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:23:59,345 (deepspeed_trainer:228) INFO: 41epoch:train:6001-6100batch: iter_time=1.114e-04, loss_ctc=70.858, loss_att=49.093, acc=0.728, loss=55.607, grad_norm=6.019, loss_scale=1.000, learning_rate=6.293e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:24:36,218 (deepspeed_trainer:228) INFO: 41epoch:train:6101-6200batch: iter_time=1.129e-04, loss_ctc=68.829, loss_att=50.812, acc=0.713, loss=56.210, grad_norm=6.374, loss_scale=1.000, learning_rate=6.292e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 18:25:13,180 (deepspeed_trainer:228) INFO: 41epoch:train:6201-6300batch: iter_time=1.136e-04, loss_ctc=66.496, loss_att=52.067, acc=0.725, loss=56.409, grad_norm=5.675, loss_scale=1.000, learning_rate=6.292e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:25:49,751 (deepspeed_trainer:228) INFO: 41epoch:train:6301-6400batch: iter_time=1.143e-04, loss_ctc=73.686, loss_att=50.516, acc=0.729, loss=57.477, grad_norm=6.489, loss_scale=1.000, learning_rate=6.291e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:26:26,725 (deepspeed_trainer:228) INFO: 41epoch:train:6401-6500batch: iter_time=1.155e-04, loss_ctc=77.821, loss_att=59.299, acc=0.711, loss=64.858, grad_norm=7.525, loss_scale=1.000, learning_rate=6.291e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:27:03,635 (deepspeed_trainer:228) INFO: 41epoch:train:6501-6600batch: iter_time=1.120e-04, loss_ctc=68.364, loss_att=56.759, acc=0.720, loss=60.232, grad_norm=6.024, loss_scale=1.000, learning_rate=6.290e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:27:40,424 (deepspeed_trainer:228) INFO: 41epoch:train:6601-6700batch: iter_time=1.094e-04, loss_ctc=63.831, loss_att=46.239, acc=0.739, loss=51.525, grad_norm=5.994, loss_scale=1.000, learning_rate=6.290e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:28:17,034 (deepspeed_trainer:228) INFO: 41epoch:train:6701-6800batch: iter_time=1.083e-04, loss_ctc=61.929, loss_att=45.279, acc=0.739, loss=50.270, grad_norm=5.823, loss_scale=1.000, learning_rate=6.289e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 18:28:53,769 (deepspeed_trainer:228) INFO: 41epoch:train:6801-6900batch: iter_time=1.106e-04, loss_ctc=64.855, loss_att=48.608, acc=0.734, loss=53.484, grad_norm=5.622, loss_scale=1.000, learning_rate=6.289e-05, step_time=0.367 [2024-12-10 18:29:30,566] [INFO] [logging.py:129:log_dist] [Rank 0] step=607000, skipped=0, lr=[np.float64(6.28797661048084e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:29:30,566] [INFO] [timer.py:264:stop] epoch=0/micro_step=7000/global_step=7000, RunningAvgSamplesPerSec=44.27469212767374, CurrSamplesPerSec=40.15503193891298, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:29:30,568 (deepspeed_trainer:228) INFO: 41epoch:train:6901-7000batch: iter_time=1.111e-04, loss_ctc=77.014, loss_att=56.744, acc=0.718, loss=62.818, grad_norm=6.538, loss_scale=1.000, learning_rate=6.288e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:30:07,526 (deepspeed_trainer:228) INFO: 41epoch:train:7001-7100batch: iter_time=1.124e-04, loss_ctc=73.285, loss_att=53.139, acc=0.736, loss=59.196, grad_norm=5.530, loss_scale=1.000, learning_rate=6.288e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:30:44,498 (deepspeed_trainer:228) INFO: 41epoch:train:7101-7200batch: iter_time=1.149e-04, loss_ctc=73.216, loss_att=53.574, acc=0.738, loss=59.467, grad_norm=6.221, loss_scale=1.000, learning_rate=6.287e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:31:21,260 (deepspeed_trainer:228) INFO: 41epoch:train:7201-7300batch: iter_time=1.149e-04, loss_ctc=66.103, loss_att=48.956, acc=0.737, loss=54.102, grad_norm=5.743, loss_scale=1.000, learning_rate=6.287e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:31:57,848 (deepspeed_trainer:228) INFO: 41epoch:train:7301-7400batch: iter_time=1.123e-04, loss_ctc=66.569, loss_att=46.659, acc=0.737, loss=52.638, grad_norm=6.045, loss_scale=1.000, learning_rate=6.286e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:32:34,148 (deepspeed_trainer:228) INFO: 41epoch:train:7401-7500batch: iter_time=1.075e-04, loss_ctc=77.899, loss_att=58.236, acc=0.715, loss=64.143, grad_norm=6.224, loss_scale=1.000, learning_rate=6.286e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 18:32:37,811 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-018:0/16] 2024-12-10 18:33:03,781 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 18:33:18,990 (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: ) [cnode7-018:0/16] 2024-12-10 18:33:18,990 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 18:33:18,992 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 18:33:45,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:46,968] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:47,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:47,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:47,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,390] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,842] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:33:48,849] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:34,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:34,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:34,783] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:36,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:35,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:36,699] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:36,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,312] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,170] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:34:37,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:22,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:22,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:23,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:24,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:24,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:25,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:25,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:25,700] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:26,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:26,398] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:26,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:26,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:26,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:27,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:27,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:35:27,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:10,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:11,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:11,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:12,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:13,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:13,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:13,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:14,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:14,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:14,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:15,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:15,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:15,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:15,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:16,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:36:21,873] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 18:37:24,649 (deepspeed_trainer:228) INFO: 41epoch:train:7501-7600batch: iter_time=2.532, loss_ctc=71.761, loss_att=59.709, acc=0.728, loss=63.322, grad_norm=5.750, loss_scale=1.000, learning_rate=6.285e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 18:38:01,722 (deepspeed_trainer:228) INFO: 41epoch:train:7601-7700batch: iter_time=1.443e-04, loss_ctc=70.504, loss_att=54.531, acc=0.733, loss=59.335, grad_norm=6.138, loss_scale=1.000, learning_rate=6.285e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 18:38:38,974 (deepspeed_trainer:228) INFO: 41epoch:train:7701-7800batch: iter_time=1.320e-04, loss_ctc=77.653, loss_att=58.597, acc=0.725, loss=64.293, grad_norm=7.255, loss_scale=1.000, learning_rate=6.284e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 18:39:15,779 (deepspeed_trainer:228) INFO: 41epoch:train:7801-7900batch: iter_time=1.283e-04, loss_ctc=66.114, loss_att=51.255, acc=0.731, loss=55.679, grad_norm=6.141, loss_scale=1.000, learning_rate=6.284e-05, step_time=0.367 [2024-12-10 18:39:52,689] [INFO] [logging.py:129:log_dist] [Rank 0] step=608000, skipped=0, lr=[np.float64(6.282803457598398e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:39:52,690] [INFO] [timer.py:264:stop] epoch=0/micro_step=8000/global_step=8000, RunningAvgSamplesPerSec=44.29562993393112, CurrSamplesPerSec=43.54949388548393, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:39:52,692 (deepspeed_trainer:228) INFO: 41epoch:train:7901-8000batch: iter_time=1.252e-04, loss_ctc=74.216, loss_att=51.717, acc=0.731, loss=58.476, grad_norm=6.567, loss_scale=1.000, learning_rate=6.283e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:40:29,181 (deepspeed_trainer:228) INFO: 41epoch:train:8001-8100batch: iter_time=1.187e-04, loss_ctc=62.219, loss_att=45.753, acc=0.727, loss=50.672, grad_norm=5.657, loss_scale=1.000, learning_rate=6.283e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:41:06,131 (deepspeed_trainer:228) INFO: 41epoch:train:8101-8200batch: iter_time=1.212e-04, loss_ctc=66.199, loss_att=49.766, acc=0.731, loss=54.669, grad_norm=5.772, loss_scale=1.000, learning_rate=6.282e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:41:43,438 (deepspeed_trainer:228) INFO: 41epoch:train:8201-8300batch: iter_time=1.187e-04, loss_ctc=76.064, loss_att=55.660, acc=0.733, loss=61.785, grad_norm=6.592, loss_scale=1.000, learning_rate=6.282e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 18:42:20,526 (deepspeed_trainer:228) INFO: 41epoch:train:8301-8400batch: iter_time=1.162e-04, loss_ctc=75.451, loss_att=62.174, acc=0.709, loss=66.164, grad_norm=7.356, loss_scale=1.000, learning_rate=6.281e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 18:42:57,439 (deepspeed_trainer:228) INFO: 41epoch:train:8401-8500batch: iter_time=1.187e-04, loss_ctc=66.399, loss_att=54.164, acc=0.736, loss=57.832, grad_norm=5.813, loss_scale=1.000, learning_rate=6.280e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:43:35,203 (deepspeed_trainer:228) INFO: 41epoch:train:8501-8600batch: iter_time=1.138e-04, loss_ctc=65.493, loss_att=48.948, acc=0.733, loss=53.903, grad_norm=6.057, loss_scale=1.000, learning_rate=6.280e-05, step_time=0.377 [cnode7-018:0/16] 2024-12-10 18:44:11,616 (deepspeed_trainer:228) INFO: 41epoch:train:8601-8700batch: iter_time=1.113e-04, loss_ctc=58.631, loss_att=43.056, acc=0.744, loss=47.742, grad_norm=5.124, loss_scale=1.000, learning_rate=6.279e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:44:48,321 (deepspeed_trainer:228) INFO: 41epoch:train:8701-8800batch: iter_time=1.142e-04, loss_ctc=74.329, loss_att=52.647, acc=0.738, loss=59.166, grad_norm=6.297, loss_scale=1.000, learning_rate=6.279e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 18:45:25,008 (deepspeed_trainer:228) INFO: 41epoch:train:8801-8900batch: iter_time=1.135e-04, loss_ctc=73.798, loss_att=57.382, acc=0.719, loss=62.309, grad_norm=6.586, loss_scale=1.000, learning_rate=6.278e-05, step_time=0.367 [2024-12-10 18:46:01,842] [INFO] [logging.py:129:log_dist] [Rank 0] step=609000, skipped=0, lr=[np.float64(6.277643051687227e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:46:01,842] [INFO] [timer.py:264:stop] epoch=0/micro_step=9000/global_step=9000, RunningAvgSamplesPerSec=44.31305045384432, CurrSamplesPerSec=44.84606017281907, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:46:01,844 (deepspeed_trainer:228) INFO: 41epoch:train:8901-9000batch: iter_time=1.118e-04, loss_ctc=76.341, loss_att=55.443, acc=0.743, loss=61.723, grad_norm=6.865, loss_scale=1.000, learning_rate=6.278e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 18:46:38,353 (deepspeed_trainer:228) INFO: 41epoch:train:9001-9100batch: iter_time=1.130e-04, loss_ctc=65.760, loss_att=50.756, acc=0.747, loss=55.239, grad_norm=5.551, loss_scale=1.000, learning_rate=6.277e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:47:14,741 (deepspeed_trainer:228) INFO: 41epoch:train:9101-9200batch: iter_time=1.109e-04, loss_ctc=66.379, loss_att=48.507, acc=0.735, loss=53.865, grad_norm=5.987, loss_scale=1.000, learning_rate=6.277e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:47:51,394 (deepspeed_trainer:228) INFO: 41epoch:train:9201-9300batch: iter_time=1.112e-04, loss_ctc=69.336, loss_att=49.241, acc=0.742, loss=55.256, grad_norm=6.057, loss_scale=1.000, learning_rate=6.276e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 18:48:22,833 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-018:0/16] 2024-12-10 18:48:50,379 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 18:49:07,117 (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: ) [cnode7-018:0/16] 2024-12-10 18:49:07,117 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 18:49:07,120 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 18:49:31,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:32,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:32,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:32,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:32,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,270] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:34,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:33,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:34,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:34,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:49:39,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:20,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:20,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,653] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:22,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:21,877] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:22,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:22,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:22,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:23,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:23,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:50:25,790] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:07,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:08,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:09,162] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:09,730] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:09,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:10,049] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:10,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:10,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:10,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,243] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:11,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:13,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:54,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:56,388] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:56,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:57,548] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:58,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:58,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:58,751] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:58,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:59,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:59,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:59,955] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:51:59,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:52:00,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:52:00,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:52:01,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 18:52:01,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 18:52:37,692 (deepspeed_trainer:228) INFO: 41epoch:train:9301-9400batch: iter_time=2.491, loss_ctc=80.619, loss_att=61.145, acc=0.711, loss=67.022, grad_norm=6.699, loss_scale=1.000, learning_rate=6.276e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 18:53:14,177 (deepspeed_trainer:228) INFO: 41epoch:train:9401-9500batch: iter_time=1.063e-04, loss_ctc=64.779, loss_att=52.284, acc=0.733, loss=56.014, grad_norm=5.605, loss_scale=1.000, learning_rate=6.275e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 18:53:51,260 (deepspeed_trainer:228) INFO: 41epoch:train:9501-9600batch: iter_time=1.169e-04, loss_ctc=83.082, loss_att=62.207, acc=0.721, loss=68.501, grad_norm=7.324, loss_scale=1.000, learning_rate=6.275e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 18:54:27,911 (deepspeed_trainer:228) INFO: 41epoch:train:9601-9700batch: iter_time=1.165e-04, loss_ctc=65.860, loss_att=53.565, acc=0.721, loss=57.265, grad_norm=5.740, loss_scale=1.000, learning_rate=6.274e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 18:55:03,342 (deepspeed_trainer:228) INFO: 41epoch:train:9701-9800batch: iter_time=1.096e-04, loss_ctc=71.096, loss_att=53.332, acc=0.717, loss=58.676, grad_norm=5.876, loss_scale=1.000, learning_rate=6.274e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 18:55:39,664 (deepspeed_trainer:228) INFO: 41epoch:train:9801-9900batch: iter_time=1.124e-04, loss_ctc=70.806, loss_att=48.104, acc=0.725, loss=54.911, grad_norm=6.274, loss_scale=1.000, learning_rate=6.273e-05, step_time=0.363 [2024-12-10 18:56:15,984] [INFO] [logging.py:129:log_dist] [Rank 0] step=610000, skipped=0, lr=[np.float64(6.272495340484284e-05)], mom=[[0.9, 0.98]] [2024-12-10 18:56:15,985] [INFO] [timer.py:264:stop] epoch=0/micro_step=10000/global_step=10000, RunningAvgSamplesPerSec=44.37985597201528, CurrSamplesPerSec=42.9636084489489, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 18:56:15,986 (deepspeed_trainer:228) INFO: 41epoch:train:9901-10000batch: iter_time=1.119e-04, loss_ctc=61.118, loss_att=46.561, acc=0.728, loss=50.932, grad_norm=5.292, loss_scale=1.000, learning_rate=6.273e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 18:56:52,912 (deepspeed_trainer:228) INFO: 41epoch:train:10001-10100batch: iter_time=1.114e-04, loss_ctc=73.623, loss_att=55.584, acc=0.723, loss=61.016, grad_norm=5.850, loss_scale=1.000, learning_rate=6.272e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 18:57:29,254 (deepspeed_trainer:228) INFO: 41epoch:train:10101-10200batch: iter_time=1.099e-04, loss_ctc=66.775, loss_att=48.906, acc=0.723, loss=54.268, grad_norm=6.760, loss_scale=1.000, learning_rate=6.272e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 18:58:06,140 (deepspeed_trainer:228) INFO: 41epoch:train:10201-10300batch: iter_time=1.115e-04, loss_ctc=80.110, loss_att=63.198, acc=0.713, loss=68.291, grad_norm=7.589, loss_scale=1.000, learning_rate=6.271e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 18:58:42,606 (deepspeed_trainer:228) INFO: 41epoch:train:10301-10400batch: iter_time=1.109e-04, loss_ctc=63.734, loss_att=49.928, acc=0.736, loss=54.073, grad_norm=5.837, loss_scale=1.000, learning_rate=6.271e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:59:18,984 (deepspeed_trainer:228) INFO: 41epoch:train:10401-10500batch: iter_time=1.099e-04, loss_ctc=66.236, loss_att=49.408, acc=0.736, loss=54.446, grad_norm=5.700, loss_scale=1.000, learning_rate=6.270e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 18:59:55,356 (deepspeed_trainer:228) INFO: 41epoch:train:10501-10600batch: iter_time=1.095e-04, loss_ctc=61.623, loss_att=45.922, acc=0.739, loss=50.648, grad_norm=5.561, loss_scale=1.000, learning_rate=6.270e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 19:00:31,776 (deepspeed_trainer:228) INFO: 41epoch:train:10601-10700batch: iter_time=1.121e-04, loss_ctc=74.332, loss_att=50.321, acc=0.735, loss=57.551, grad_norm=6.632, loss_scale=1.000, learning_rate=6.269e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:01:08,116 (deepspeed_trainer:228) INFO: 41epoch:train:10701-10800batch: iter_time=1.103e-04, loss_ctc=71.137, loss_att=54.779, acc=0.717, loss=59.717, grad_norm=5.992, loss_scale=1.000, learning_rate=6.269e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 19:01:44,715 (deepspeed_trainer:228) INFO: 41epoch:train:10801-10900batch: iter_time=1.098e-04, loss_ctc=79.820, loss_att=55.232, acc=0.749, loss=62.628, grad_norm=6.032, loss_scale=1.000, learning_rate=6.268e-05, step_time=0.366 [2024-12-10 19:02:20,859] [INFO] [logging.py:129:log_dist] [Rank 0] step=611000, skipped=0, lr=[np.float64(6.267360272026034e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:02:20,860] [INFO] [timer.py:264:stop] epoch=0/micro_step=11000/global_step=11000, RunningAvgSamplesPerSec=44.437448821287795, CurrSamplesPerSec=43.38120294931194, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:02:20,862 (deepspeed_trainer:228) INFO: 41epoch:train:10901-11000batch: iter_time=1.102e-04, loss_ctc=61.601, loss_att=49.054, acc=0.733, loss=52.814, grad_norm=5.773, loss_scale=1.000, learning_rate=6.268e-05, step_time=0.361 [cnode7-018:0/16] 2024-12-10 19:02:57,148 (deepspeed_trainer:228) INFO: 41epoch:train:11001-11100batch: iter_time=1.097e-04, loss_ctc=66.218, loss_att=46.197, acc=0.737, loss=52.180, grad_norm=5.848, loss_scale=1.000, learning_rate=6.267e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 19:03:33,594 (deepspeed_trainer:228) INFO: 41epoch:train:11101-11200batch: iter_time=1.093e-04, loss_ctc=77.070, loss_att=54.801, acc=0.729, loss=61.473, grad_norm=5.916, loss_scale=1.000, learning_rate=6.267e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:03:55,329 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-018:0/16] 2024-12-10 19:04:22,490 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 19:04:38,983 (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: ) [cnode7-018:0/16] 2024-12-10 19:04:38,983 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 19:04:38,985 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 19:05:04,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:04,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:05,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,820] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:06,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:07,186] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:07,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:07,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:08,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:08,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:08,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:52,776] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:54,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:54,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:54,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:55,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:56,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:56,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:56,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:56,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:56,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:05:57,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:40,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:43,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:43,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:43,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:43,899] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:44,259] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:43,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:44,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:44,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:44,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,145] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,198] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:06:45,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:28,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:31,113] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:31,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,872] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:32,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:33,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:33,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:33,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:33,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:33,791] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:07:38,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 19:08:23,230 (deepspeed_trainer:228) INFO: 41epoch:train:11201-11300batch: iter_time=2.474, loss_ctc=78.825, loss_att=64.412, acc=0.713, loss=68.724, grad_norm=7.105, loss_scale=1.000, learning_rate=6.266e-05, step_time=0.422 [cnode7-018:0/16] 2024-12-10 19:08:59,965 (deepspeed_trainer:228) INFO: 41epoch:train:11301-11400batch: iter_time=1.086e-04, loss_ctc=63.243, loss_att=46.282, acc=0.742, loss=51.374, grad_norm=5.210, loss_scale=1.000, learning_rate=6.266e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:09:36,894 (deepspeed_trainer:228) INFO: 41epoch:train:11401-11500batch: iter_time=1.065e-04, loss_ctc=78.066, loss_att=61.907, acc=0.725, loss=66.794, grad_norm=7.701, loss_scale=1.000, learning_rate=6.265e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:10:13,327 (deepspeed_trainer:228) INFO: 41epoch:train:11501-11600batch: iter_time=1.108e-04, loss_ctc=63.746, loss_att=49.189, acc=0.736, loss=53.532, grad_norm=5.331, loss_scale=1.000, learning_rate=6.265e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:10:49,738 (deepspeed_trainer:228) INFO: 41epoch:train:11601-11700batch: iter_time=1.115e-04, loss_ctc=72.922, loss_att=54.273, acc=0.728, loss=59.847, grad_norm=5.958, loss_scale=1.000, learning_rate=6.264e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:11:26,108 (deepspeed_trainer:228) INFO: 41epoch:train:11701-11800batch: iter_time=1.089e-04, loss_ctc=69.626, loss_att=47.426, acc=0.732, loss=54.084, grad_norm=6.046, loss_scale=1.000, learning_rate=6.264e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 19:12:02,432 (deepspeed_trainer:228) INFO: 41epoch:train:11801-11900batch: iter_time=1.166e-04, loss_ctc=62.141, loss_att=50.341, acc=0.729, loss=53.925, grad_norm=5.011, loss_scale=1.000, learning_rate=6.263e-05, step_time=0.362 [2024-12-10 19:12:38,985] [INFO] [logging.py:129:log_dist] [Rank 0] step=612000, skipped=0, lr=[np.float64(6.262237794646236e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:12:38,985] [INFO] [timer.py:264:stop] epoch=0/micro_step=12000/global_step=12000, RunningAvgSamplesPerSec=44.423936526223656, CurrSamplesPerSec=46.67464755618426, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:12:38,987 (deepspeed_trainer:228) INFO: 41epoch:train:11901-12000batch: iter_time=1.125e-04, loss_ctc=71.956, loss_att=53.290, acc=0.729, loss=58.870, grad_norm=6.072, loss_scale=1.000, learning_rate=6.262e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 19:13:16,250 (deepspeed_trainer:228) INFO: 41epoch:train:12001-12100batch: iter_time=1.125e-04, loss_ctc=76.751, loss_att=55.307, acc=0.716, loss=61.721, grad_norm=7.333, loss_scale=1.000, learning_rate=6.262e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:13:53,415 (deepspeed_trainer:228) INFO: 41epoch:train:12101-12200batch: iter_time=1.123e-04, loss_ctc=75.875, loss_att=61.151, acc=0.731, loss=65.556, grad_norm=6.144, loss_scale=1.000, learning_rate=6.261e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 19:14:29,913 (deepspeed_trainer:228) INFO: 41epoch:train:12201-12300batch: iter_time=1.129e-04, loss_ctc=62.171, loss_att=51.232, acc=0.734, loss=54.502, grad_norm=6.082, loss_scale=1.000, learning_rate=6.261e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:15:06,226 (deepspeed_trainer:228) INFO: 41epoch:train:12301-12400batch: iter_time=1.112e-04, loss_ctc=61.263, loss_att=46.023, acc=0.744, loss=50.610, grad_norm=5.419, loss_scale=1.000, learning_rate=6.260e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 19:15:42,599 (deepspeed_trainer:228) INFO: 41epoch:train:12401-12500batch: iter_time=1.114e-04, loss_ctc=65.297, loss_att=49.418, acc=0.740, loss=54.173, grad_norm=5.058, loss_scale=1.000, learning_rate=6.260e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 19:16:19,360 (deepspeed_trainer:228) INFO: 41epoch:train:12501-12600batch: iter_time=1.114e-04, loss_ctc=74.267, loss_att=52.493, acc=0.732, loss=59.011, grad_norm=6.678, loss_scale=1.000, learning_rate=6.259e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:16:55,761 (deepspeed_trainer:228) INFO: 41epoch:train:12601-12700batch: iter_time=1.116e-04, loss_ctc=70.529, loss_att=53.355, acc=0.734, loss=58.514, grad_norm=6.348, loss_scale=1.000, learning_rate=6.259e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:17:32,696 (deepspeed_trainer:228) INFO: 41epoch:train:12701-12800batch: iter_time=1.124e-04, loss_ctc=80.117, loss_att=57.504, acc=0.745, loss=64.282, grad_norm=6.134, loss_scale=1.000, learning_rate=6.258e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:18:09,268 (deepspeed_trainer:228) INFO: 41epoch:train:12801-12900batch: iter_time=1.118e-04, loss_ctc=63.111, loss_att=49.119, acc=0.740, loss=53.296, grad_norm=5.754, loss_scale=1.000, learning_rate=6.258e-05, step_time=0.365 [2024-12-10 19:18:46,051] [INFO] [logging.py:129:log_dist] [Rank 0] step=613000, skipped=0, lr=[np.float64(6.257127856973765e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:18:46,051] [INFO] [timer.py:264:stop] epoch=0/micro_step=13000/global_step=13000, RunningAvgSamplesPerSec=44.44757957131839, CurrSamplesPerSec=47.27400569641526, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:18:46,053 (deepspeed_trainer:228) INFO: 41epoch:train:12901-13000batch: iter_time=1.142e-04, loss_ctc=62.696, loss_att=44.701, acc=0.738, loss=50.096, grad_norm=5.682, loss_scale=1.000, learning_rate=6.257e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 19:19:22,985 (deepspeed_trainer:228) INFO: 41epoch:train:13001-13100batch: iter_time=1.135e-04, loss_ctc=80.142, loss_att=58.546, acc=0.725, loss=64.993, grad_norm=6.177, loss_scale=1.000, learning_rate=6.257e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:19:35,339 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-018:0/16] 2024-12-10 19:20:02,721 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 19:20:19,136 (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: ) [cnode7-018:0/16] 2024-12-10 19:20:19,136 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 19:20:19,138 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 19:20:45,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:45,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:45,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:46,354] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:46,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,528] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,589] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:47,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:48,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:48,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:48,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:48,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:20:48,791] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:33,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:33,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:34,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:34,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:35,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:36,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:36,173] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:37,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:38,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:38,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:21:38,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:21,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:21,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:21,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:22,534] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:23,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:24,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:24,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:25,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:25,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:25,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:26,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:26,329] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:27,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:27,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:27,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:22:30,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:08,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:09,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:10,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:12,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:12,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:13,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:13,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:13,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:13,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:13,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:14,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:15,113] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:15,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:16,805] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:17,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:23:20,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 19:24:15,278 (deepspeed_trainer:228) INFO: 41epoch:train:13101-13200batch: iter_time=2.498, loss_ctc=70.472, loss_att=56.608, acc=0.727, loss=60.777, grad_norm=5.649, loss_scale=1.000, learning_rate=6.256e-05, step_time=0.425 [cnode7-018:0/16] 2024-12-10 19:24:52,406 (deepspeed_trainer:228) INFO: 41epoch:train:13201-13300batch: iter_time=1.103e-04, loss_ctc=67.371, loss_att=49.798, acc=0.737, loss=55.064, grad_norm=6.016, loss_scale=1.000, learning_rate=6.256e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:25:29,668 (deepspeed_trainer:228) INFO: 41epoch:train:13301-13400batch: iter_time=1.103e-04, loss_ctc=80.826, loss_att=64.806, acc=0.722, loss=69.593, grad_norm=6.911, loss_scale=1.000, learning_rate=6.255e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 19:26:06,381 (deepspeed_trainer:228) INFO: 41epoch:train:13401-13500batch: iter_time=1.125e-04, loss_ctc=66.130, loss_att=49.681, acc=0.740, loss=54.609, grad_norm=5.532, loss_scale=1.000, learning_rate=6.255e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:26:43,189 (deepspeed_trainer:228) INFO: 41epoch:train:13501-13600batch: iter_time=1.119e-04, loss_ctc=70.332, loss_att=49.388, acc=0.731, loss=55.690, grad_norm=5.851, loss_scale=1.000, learning_rate=6.254e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 19:27:19,911 (deepspeed_trainer:228) INFO: 41epoch:train:13601-13700batch: iter_time=1.101e-04, loss_ctc=67.844, loss_att=50.246, acc=0.721, loss=55.552, grad_norm=6.236, loss_scale=1.000, learning_rate=6.254e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:27:56,639 (deepspeed_trainer:228) INFO: 41epoch:train:13701-13800batch: iter_time=1.217e-04, loss_ctc=66.096, loss_att=52.063, acc=0.731, loss=56.285, grad_norm=6.063, loss_scale=1.000, learning_rate=6.253e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:28:33,314 (deepspeed_trainer:228) INFO: 41epoch:train:13801-13900batch: iter_time=1.098e-04, loss_ctc=72.156, loss_att=51.158, acc=0.734, loss=57.454, grad_norm=6.914, loss_scale=1.000, learning_rate=6.253e-05, step_time=0.366 [2024-12-10 19:29:10,354] [INFO] [logging.py:129:log_dist] [Rank 0] step=614000, skipped=0, lr=[np.float64(6.252030407930445e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:29:10,354] [INFO] [timer.py:264:stop] epoch=0/micro_step=14000/global_step=14000, RunningAvgSamplesPerSec=44.40495310078583, CurrSamplesPerSec=43.75598174751613, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:29:10,356 (deepspeed_trainer:228) INFO: 41epoch:train:13901-14000batch: iter_time=1.105e-04, loss_ctc=77.217, loss_att=59.416, acc=0.716, loss=64.768, grad_norm=8.020, loss_scale=1.000, learning_rate=6.252e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 19:29:47,412 (deepspeed_trainer:228) INFO: 41epoch:train:14001-14100batch: iter_time=1.211e-04, loss_ctc=67.876, loss_att=55.919, acc=0.731, loss=59.508, grad_norm=6.128, loss_scale=1.000, learning_rate=6.252e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 19:30:23,871 (deepspeed_trainer:228) INFO: 41epoch:train:14101-14200batch: iter_time=1.125e-04, loss_ctc=63.111, loss_att=47.992, acc=0.737, loss=52.550, grad_norm=6.012, loss_scale=1.000, learning_rate=6.251e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:31:00,616 (deepspeed_trainer:228) INFO: 41epoch:train:14201-14300batch: iter_time=1.136e-04, loss_ctc=61.240, loss_att=45.392, acc=0.746, loss=50.160, grad_norm=4.912, loss_scale=1.000, learning_rate=6.251e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:31:37,155 (deepspeed_trainer:228) INFO: 41epoch:train:14301-14400batch: iter_time=1.095e-04, loss_ctc=64.208, loss_att=48.080, acc=0.744, loss=52.920, grad_norm=5.176, loss_scale=1.000, learning_rate=6.250e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 19:32:13,672 (deepspeed_trainer:228) INFO: 41epoch:train:14401-14500batch: iter_time=1.106e-04, loss_ctc=76.279, loss_att=56.574, acc=0.720, loss=62.498, grad_norm=7.065, loss_scale=1.000, learning_rate=6.250e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:32:50,369 (deepspeed_trainer:228) INFO: 41epoch:train:14501-14600batch: iter_time=1.104e-04, loss_ctc=73.256, loss_att=53.243, acc=0.741, loss=59.243, grad_norm=5.950, loss_scale=1.000, learning_rate=6.249e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:33:27,509 (deepspeed_trainer:228) INFO: 41epoch:train:14601-14700batch: iter_time=1.107e-04, loss_ctc=72.418, loss_att=54.969, acc=0.741, loss=60.179, grad_norm=6.048, loss_scale=1.000, learning_rate=6.249e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:34:04,233 (deepspeed_trainer:228) INFO: 41epoch:train:14701-14800batch: iter_time=1.119e-04, loss_ctc=65.878, loss_att=49.002, acc=0.742, loss=54.060, grad_norm=5.338, loss_scale=1.000, learning_rate=6.248e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:34:40,811 (deepspeed_trainer:228) INFO: 41epoch:train:14801-14900batch: iter_time=1.143e-04, loss_ctc=65.794, loss_att=47.970, acc=0.736, loss=53.324, grad_norm=5.516, loss_scale=1.000, learning_rate=6.248e-05, step_time=0.365 [2024-12-10 19:35:17,249] [INFO] [logging.py:129:log_dist] [Rank 0] step=615000, skipped=0, lr=[np.float64(6.246945396728912e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:35:17,250] [INFO] [timer.py:264:stop] epoch=0/micro_step=15000/global_step=15000, RunningAvgSamplesPerSec=44.4288731050137, CurrSamplesPerSec=46.90145626010551, MemAllocated=1.77GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:35:17,252 (deepspeed_trainer:228) INFO: 41epoch:train:14901-15000batch: iter_time=1.097e-04, loss_ctc=77.228, loss_att=57.938, acc=0.722, loss=63.725, grad_norm=6.635, loss_scale=1.000, learning_rate=6.247e-05, step_time=0.365 [2024-12-10 19:35:31,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:31,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:31,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:31,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,037] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,254] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,610] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,899] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:33,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:32,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:46,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:47,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:47,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:48,740] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:48,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:48,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:48,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:49,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:50,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:50,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:35:50,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:01,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:03,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:03,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:04,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:04,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:04,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:04,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:05,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:05,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:05,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:05,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:06,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:06,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:07,115] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:07,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:07,242] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:16,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:18,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:19,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:19,740] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:20,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:20,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:20,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:21,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:21,159] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:22,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:21,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:21,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:22,375] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:22,755] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:23,089] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:23,396] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:36:35,829] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 41 is about to be saved! [2024-12-10 19:36:35,862] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/mp_rank_00_model_states.pt [2024-12-10 19:36:35,862] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/mp_rank_00_model_states.pt... [2024-12-10 19:36:37,852] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/mp_rank_00_model_states.pt. [2024-12-10 19:36:37,659] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,009] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,010] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,660] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,011] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,011] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,661] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,662] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,662] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,016] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,017] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,017] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,018] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,667] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,668] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-10 19:36:37,668] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-10 19:36:38,804] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,804] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,804] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:38,830] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,830] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,830] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:38,485] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,486] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,486] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:38,889] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,889] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,889] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:38,899] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,902] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,902] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,902] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:38,635] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-10 19:36:38,635] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-12-10 19:36:38,635] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,079] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,079] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,170] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,170] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,170] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,301] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,301] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,301] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,192] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,192] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,192] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,563] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,563] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,563] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,216] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,216] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,216] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,820] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,820] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,821] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,863] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,863] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,863] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,869] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,870] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,870] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [2024-12-10 19:36:39,881] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-10 19:36:39,882] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_41/41/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-12-10 19:36:39,882] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 41 is ready now! [cnode7-018:0/16] 2024-12-10 19:36:40,254 (deepspeed_trainer:158) INFO: 41epoch results: [train] iter_time=0.132, loss_ctc=70.486, loss_att=52.914, acc=0.729, loss=58.186, grad_norm=6.189, loss_scale=1.000, learning_rate=6.286e-05, step_time=0.370, time=2 hours, 5 minutes and 42.59 seconds, total_count=615041, gpu_max_cached_mem_GB=28.652, [valid] loss_ctc=3.875, cer_ctc=0.071, loss_att=5.938, acc=0.866, cer=0.225, wer=0.812, loss=5.344, time=1 minute and 10.29 seconds, total_count=41, gpu_max_cached_mem_GB=28.652 [cnode7-018:0/16] 2024-12-10 19:36:42,429 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-018:0/16] 2024-12-10 19:37:08,585 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 19:37:24,149 (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: ) [cnode7-018:0/16] 2024-12-10 19:37:24,149 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 19:37:24,151 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 19:37:44,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:44,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:45,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:45,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:45,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:45,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:45,459] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,082] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,942] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:46,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:37:47,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:31,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:32,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:32,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:32,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:32,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:33,330] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:33,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:33,835] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:35,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:34,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:38:38,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:18,650] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:19,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:20,652] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:21,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:21,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:21,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:21,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:22,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:23,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:39:28,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:06,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:07,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:08,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:08,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:09,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:09,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:09,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:10,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:10,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:11,859] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:40:18,109] [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): [cnode7-018:0/16] 2024-12-10 19:41:20,877 (deepspeed_trainer:228) INFO: 42epoch:train:1-100batch: iter_time=2.348, loss_ctc=71.468, loss_att=52.226, acc=0.728, loss=57.990, grad_norm=6.269, loss_scale=1.000, learning_rate=6.247e-05, step_time=0.441 [cnode7-018:0/16] 2024-12-10 19:41:57,858 (deepspeed_trainer:228) INFO: 42epoch:train:101-200batch: iter_time=1.056e-04, loss_ctc=63.151, loss_att=48.296, acc=0.720, loss=52.746, grad_norm=5.800, loss_scale=1.000, learning_rate=6.246e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:42:34,696 (deepspeed_trainer:228) INFO: 42epoch:train:201-300batch: iter_time=1.098e-04, loss_ctc=66.567, loss_att=48.047, acc=0.728, loss=53.599, grad_norm=6.117, loss_scale=1.000, learning_rate=6.246e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 19:43:11,919 (deepspeed_trainer:228) INFO: 42epoch:train:301-400batch: iter_time=1.117e-04, loss_ctc=69.493, loss_att=51.968, acc=0.727, loss=57.214, grad_norm=6.032, loss_scale=1.000, learning_rate=6.245e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 19:43:48,718 (deepspeed_trainer:228) INFO: 42epoch:train:401-500batch: iter_time=1.141e-04, loss_ctc=69.414, loss_att=50.918, acc=0.727, loss=56.458, grad_norm=6.302, loss_scale=1.000, learning_rate=6.245e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 19:44:25,926 (deepspeed_trainer:228) INFO: 42epoch:train:501-600batch: iter_time=1.105e-04, loss_ctc=72.513, loss_att=52.602, acc=0.739, loss=58.583, grad_norm=6.399, loss_scale=1.000, learning_rate=6.244e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:45:03,040 (deepspeed_trainer:228) INFO: 42epoch:train:601-700batch: iter_time=1.097e-04, loss_ctc=67.053, loss_att=50.864, acc=0.735, loss=55.727, grad_norm=6.560, loss_scale=1.000, learning_rate=6.244e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:45:40,919 (deepspeed_trainer:228) INFO: 42epoch:train:701-800batch: iter_time=1.095e-04, loss_ctc=66.429, loss_att=49.537, acc=0.740, loss=54.589, grad_norm=5.710, loss_scale=1.000, learning_rate=6.243e-05, step_time=0.379 [cnode7-018:0/16] 2024-12-10 19:46:18,432 (deepspeed_trainer:228) INFO: 42epoch:train:801-900batch: iter_time=1.110e-04, loss_ctc=68.294, loss_att=52.483, acc=0.732, loss=57.195, grad_norm=5.466, loss_scale=1.000, learning_rate=6.243e-05, step_time=0.375 [2024-12-10 19:46:55,153] [INFO] [logging.py:129:log_dist] [Rank 0] step=616000, skipped=0, lr=[np.float64(6.241872772870475e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:46:55,154] [INFO] [timer.py:264:stop] epoch=0/micro_step=16000/global_step=16000, RunningAvgSamplesPerSec=44.36064692971436, CurrSamplesPerSec=46.74412321924203, MemAllocated=2.04GB, MaxMemAllocated=15.7GB [cnode7-018:0/16] 2024-12-10 19:46:55,156 (deepspeed_trainer:228) INFO: 42epoch:train:901-1000batch: iter_time=1.131e-04, loss_ctc=71.558, loss_att=55.000, acc=0.707, loss=59.958, grad_norm=6.409, loss_scale=1.000, learning_rate=6.242e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:47:32,430 (deepspeed_trainer:228) INFO: 42epoch:train:1001-1100batch: iter_time=1.091e-04, loss_ctc=69.115, loss_att=53.569, acc=0.727, loss=58.221, grad_norm=6.002, loss_scale=1.000, learning_rate=6.242e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 19:48:10,020 (deepspeed_trainer:228) INFO: 42epoch:train:1101-1200batch: iter_time=1.096e-04, loss_ctc=65.293, loss_att=43.969, acc=0.749, loss=50.377, grad_norm=5.305, loss_scale=1.000, learning_rate=6.241e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-10 19:48:47,226 (deepspeed_trainer:228) INFO: 42epoch:train:1201-1300batch: iter_time=1.115e-04, loss_ctc=69.013, loss_att=51.638, acc=0.738, loss=56.836, grad_norm=5.388, loss_scale=1.000, learning_rate=6.241e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 19:49:24,002 (deepspeed_trainer:228) INFO: 42epoch:train:1301-1400batch: iter_time=1.125e-04, loss_ctc=70.284, loss_att=55.396, acc=0.725, loss=59.873, grad_norm=6.321, loss_scale=1.000, learning_rate=6.240e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 19:50:00,837 (deepspeed_trainer:228) INFO: 42epoch:train:1401-1500batch: iter_time=1.098e-04, loss_ctc=65.249, loss_att=45.246, acc=0.748, loss=51.231, grad_norm=5.282, loss_scale=1.000, learning_rate=6.240e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 19:50:37,929 (deepspeed_trainer:228) INFO: 42epoch:train:1501-1600batch: iter_time=1.099e-04, loss_ctc=72.487, loss_att=58.504, acc=0.719, loss=62.704, grad_norm=6.026, loss_scale=1.000, learning_rate=6.239e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 19:51:14,391 (deepspeed_trainer:228) INFO: 42epoch:train:1601-1700batch: iter_time=1.080e-04, loss_ctc=59.525, loss_att=40.514, acc=0.741, loss=46.208, grad_norm=5.043, loss_scale=1.000, learning_rate=6.239e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 19:51:51,393 (deepspeed_trainer:228) INFO: 42epoch:train:1701-1800batch: iter_time=1.118e-04, loss_ctc=80.388, loss_att=56.810, acc=0.721, loss=63.873, grad_norm=6.352, loss_scale=1.000, learning_rate=6.238e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:52:23,990 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-018:0/16] 2024-12-10 19:52:50,316 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 19:53:06,262 (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: ) [cnode7-018:0/16] 2024-12-10 19:53:06,262 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 19:53:06,264 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 19:53:39,993] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:40,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:53:41,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:33,523] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:34,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:34,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:34,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,311] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:35,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:36,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:36,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:36,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:36,279] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:54:36,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:20,544] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:22,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:23,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:23,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:23,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:24,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:25,000] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:25,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:55:28,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:08,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:10,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:11,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:12,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:12,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:12,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,109] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:13,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:14,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:14,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:14,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 19:56:19,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 19:56:54,190 (deepspeed_trainer:228) INFO: 42epoch:train:1801-1900batch: iter_time=2.616, loss_ctc=67.322, loss_att=50.797, acc=0.723, loss=55.759, grad_norm=6.002, loss_scale=1.000, learning_rate=6.238e-05, step_time=0.412 [2024-12-10 19:57:31,375] [INFO] [logging.py:129:log_dist] [Rank 0] step=617000, skipped=0, lr=[np.float64(6.236812486143029e-05)], mom=[[0.9, 0.98]] [2024-12-10 19:57:31,378] [INFO] [timer.py:264:stop] epoch=0/micro_step=17000/global_step=17000, RunningAvgSamplesPerSec=44.32573316259868, CurrSamplesPerSec=35.99947697955308, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 19:57:31,379 (deepspeed_trainer:228) INFO: 42epoch:train:1901-2000batch: iter_time=1.114e-04, loss_ctc=70.660, loss_att=52.398, acc=0.724, loss=57.878, grad_norm=6.076, loss_scale=1.000, learning_rate=6.237e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 19:58:07,934 (deepspeed_trainer:228) INFO: 42epoch:train:2001-2100batch: iter_time=1.096e-04, loss_ctc=65.141, loss_att=49.046, acc=0.715, loss=53.881, grad_norm=6.528, loss_scale=1.000, learning_rate=6.237e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 19:58:44,441 (deepspeed_trainer:228) INFO: 42epoch:train:2101-2200batch: iter_time=1.095e-04, loss_ctc=66.168, loss_att=50.593, acc=0.723, loss=55.275, grad_norm=6.069, loss_scale=1.000, learning_rate=6.236e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 19:59:21,335 (deepspeed_trainer:228) INFO: 42epoch:train:2201-2300batch: iter_time=1.122e-04, loss_ctc=67.447, loss_att=50.227, acc=0.723, loss=55.417, grad_norm=5.828, loss_scale=1.000, learning_rate=6.236e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 19:59:57,785 (deepspeed_trainer:228) INFO: 42epoch:train:2301-2400batch: iter_time=1.115e-04, loss_ctc=68.009, loss_att=48.593, acc=0.724, loss=54.442, grad_norm=6.318, loss_scale=1.000, learning_rate=6.235e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 20:00:34,623 (deepspeed_trainer:228) INFO: 42epoch:train:2401-2500batch: iter_time=1.126e-04, loss_ctc=75.995, loss_att=56.361, acc=0.730, loss=62.253, grad_norm=6.355, loss_scale=1.000, learning_rate=6.235e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:01:11,036 (deepspeed_trainer:228) INFO: 42epoch:train:2501-2600batch: iter_time=1.133e-04, loss_ctc=65.107, loss_att=49.684, acc=0.734, loss=54.321, grad_norm=6.857, loss_scale=1.000, learning_rate=6.234e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 20:01:47,643 (deepspeed_trainer:228) INFO: 42epoch:train:2601-2700batch: iter_time=1.138e-04, loss_ctc=68.095, loss_att=50.100, acc=0.736, loss=55.512, grad_norm=5.483, loss_scale=1.000, learning_rate=6.234e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:02:24,558 (deepspeed_trainer:228) INFO: 42epoch:train:2701-2800batch: iter_time=1.103e-04, loss_ctc=69.905, loss_att=54.158, acc=0.717, loss=58.866, grad_norm=5.908, loss_scale=1.000, learning_rate=6.233e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:03:01,214 (deepspeed_trainer:228) INFO: 42epoch:train:2801-2900batch: iter_time=1.130e-04, loss_ctc=70.221, loss_att=51.477, acc=0.719, loss=57.118, grad_norm=6.334, loss_scale=1.000, learning_rate=6.233e-05, step_time=0.366 [2024-12-10 20:03:37,596] [INFO] [logging.py:129:log_dist] [Rank 0] step=618000, skipped=0, lr=[np.float64(6.231764486618945e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:03:37,597] [INFO] [timer.py:264:stop] epoch=0/micro_step=18000/global_step=18000, RunningAvgSamplesPerSec=44.35382292881582, CurrSamplesPerSec=42.443872541607085, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:03:37,599 (deepspeed_trainer:228) INFO: 42epoch:train:2901-3000batch: iter_time=1.131e-04, loss_ctc=61.886, loss_att=44.972, acc=0.733, loss=50.057, grad_norm=5.728, loss_scale=1.000, learning_rate=6.232e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 20:04:14,278 (deepspeed_trainer:228) INFO: 42epoch:train:3001-3100batch: iter_time=1.118e-04, loss_ctc=67.168, loss_att=48.220, acc=0.736, loss=53.899, grad_norm=5.110, loss_scale=1.000, learning_rate=6.232e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 20:04:50,923 (deepspeed_trainer:228) INFO: 42epoch:train:3101-3200batch: iter_time=1.119e-04, loss_ctc=68.935, loss_att=52.298, acc=0.733, loss=57.278, grad_norm=5.795, loss_scale=1.000, learning_rate=6.231e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:05:27,582 (deepspeed_trainer:228) INFO: 42epoch:train:3201-3300batch: iter_time=1.123e-04, loss_ctc=69.949, loss_att=51.743, acc=0.732, loss=57.209, grad_norm=5.612, loss_scale=1.000, learning_rate=6.231e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:06:04,157 (deepspeed_trainer:228) INFO: 42epoch:train:3301-3400batch: iter_time=1.110e-04, loss_ctc=71.640, loss_att=54.385, acc=0.732, loss=59.534, grad_norm=5.765, loss_scale=1.000, learning_rate=6.230e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 20:06:40,776 (deepspeed_trainer:228) INFO: 42epoch:train:3401-3500batch: iter_time=1.176e-04, loss_ctc=65.008, loss_att=48.652, acc=0.724, loss=53.556, grad_norm=6.057, loss_scale=1.000, learning_rate=6.229e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:07:17,648 (deepspeed_trainer:228) INFO: 42epoch:train:3501-3600batch: iter_time=1.122e-04, loss_ctc=62.661, loss_att=44.320, acc=0.738, loss=49.865, grad_norm=5.405, loss_scale=1.000, learning_rate=6.229e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:07:54,381 (deepspeed_trainer:228) INFO: 42epoch:train:3601-3700batch: iter_time=1.139e-04, loss_ctc=73.489, loss_att=49.864, acc=0.724, loss=56.956, grad_norm=5.799, loss_scale=1.000, learning_rate=6.228e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 20:08:15,914 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-018:0/16] 2024-12-10 20:08:42,651 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 20:08:59,392 (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: ) [cnode7-018:0/16] 2024-12-10 20:08:59,392 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 20:08:59,395 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 20:09:25,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:25,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:25,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:25,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:25,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:26,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:27,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:27,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:27,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:27,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:09:27,497] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:13,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:13,477] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:13,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:14,003] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:14,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:14,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:14,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:15,926] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:16,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:17,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:10:19,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:01,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:01,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:01,648] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:02,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:02,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:03,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:03,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:03,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:03,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:03,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:04,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:04,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:04,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:04,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:08,532] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:10,792] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:49,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:49,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:49,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:49,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:50,649] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:50,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:51,697] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:52,075] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:52,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:52,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:52,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:53,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:53,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:55,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:11:59,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:12:00,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 20:12:45,241 (deepspeed_trainer:228) INFO: 42epoch:train:3701-3800batch: iter_time=2.464, loss_ctc=72.170, loss_att=57.436, acc=0.717, loss=61.856, grad_norm=5.997, loss_scale=1.000, learning_rate=6.228e-05, step_time=0.445 [cnode7-018:0/16] 2024-12-10 20:13:22,222 (deepspeed_trainer:228) INFO: 42epoch:train:3801-3900batch: iter_time=1.038e-04, loss_ctc=69.122, loss_att=51.552, acc=0.721, loss=56.842, grad_norm=6.648, loss_scale=1.000, learning_rate=6.227e-05, step_time=0.369 [2024-12-10 20:13:58,858] [INFO] [logging.py:129:log_dist] [Rank 0] step=619000, skipped=0, lr=[np.float64(6.226728724653022e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:13:58,859] [INFO] [timer.py:264:stop] epoch=0/micro_step=19000/global_step=19000, RunningAvgSamplesPerSec=44.32247532979604, CurrSamplesPerSec=45.61777426619047, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:13:58,860 (deepspeed_trainer:228) INFO: 42epoch:train:3901-4000batch: iter_time=1.040e-04, loss_ctc=66.354, loss_att=47.456, acc=0.724, loss=53.140, grad_norm=6.384, loss_scale=1.000, learning_rate=6.227e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:14:35,546 (deepspeed_trainer:228) INFO: 42epoch:train:4001-4100batch: iter_time=1.093e-04, loss_ctc=67.494, loss_att=49.177, acc=0.748, loss=54.700, grad_norm=5.789, loss_scale=1.000, learning_rate=6.226e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:15:12,541 (deepspeed_trainer:228) INFO: 42epoch:train:4101-4200batch: iter_time=1.123e-04, loss_ctc=68.462, loss_att=51.370, acc=0.720, loss=56.484, grad_norm=6.582, loss_scale=1.000, learning_rate=6.226e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 20:15:48,938 (deepspeed_trainer:228) INFO: 42epoch:train:4201-4300batch: iter_time=1.167e-04, loss_ctc=63.669, loss_att=46.500, acc=0.739, loss=51.661, grad_norm=5.427, loss_scale=1.000, learning_rate=6.225e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:16:26,047 (deepspeed_trainer:228) INFO: 42epoch:train:4301-4400batch: iter_time=1.117e-04, loss_ctc=76.419, loss_att=56.129, acc=0.739, loss=62.193, grad_norm=7.005, loss_scale=1.000, learning_rate=6.225e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 20:17:03,102 (deepspeed_trainer:228) INFO: 42epoch:train:4401-4500batch: iter_time=1.114e-04, loss_ctc=67.308, loss_att=52.929, acc=0.736, loss=57.244, grad_norm=5.590, loss_scale=1.000, learning_rate=6.224e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 20:17:39,934 (deepspeed_trainer:228) INFO: 42epoch:train:4501-4600batch: iter_time=1.098e-04, loss_ctc=66.763, loss_att=50.700, acc=0.743, loss=55.511, grad_norm=5.890, loss_scale=1.000, learning_rate=6.224e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:18:16,856 (deepspeed_trainer:228) INFO: 42epoch:train:4601-4700batch: iter_time=1.134e-04, loss_ctc=67.864, loss_att=50.704, acc=0.719, loss=55.846, grad_norm=5.771, loss_scale=1.000, learning_rate=6.223e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:18:54,082 (deepspeed_trainer:228) INFO: 42epoch:train:4701-4800batch: iter_time=1.120e-04, loss_ctc=68.781, loss_att=53.053, acc=0.731, loss=57.779, grad_norm=5.590, loss_scale=1.000, learning_rate=6.223e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 20:19:31,160 (deepspeed_trainer:228) INFO: 42epoch:train:4801-4900batch: iter_time=1.101e-04, loss_ctc=62.313, loss_att=44.836, acc=0.738, loss=50.066, grad_norm=6.119, loss_scale=1.000, learning_rate=6.222e-05, step_time=0.370 [2024-12-10 20:20:08,375] [INFO] [logging.py:129:log_dist] [Rank 0] step=620000, skipped=0, lr=[np.float64(6.221705150880421e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:20:08,375] [INFO] [timer.py:264:stop] epoch=0/micro_step=20000/global_step=20000, RunningAvgSamplesPerSec=44.32638862763426, CurrSamplesPerSec=46.9193603052926, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:20:08,377 (deepspeed_trainer:228) INFO: 42epoch:train:4901-5000batch: iter_time=1.108e-04, loss_ctc=68.493, loss_att=48.098, acc=0.746, loss=54.236, grad_norm=5.643, loss_scale=1.000, learning_rate=6.222e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 20:20:45,404 (deepspeed_trainer:228) INFO: 42epoch:train:5001-5100batch: iter_time=1.108e-04, loss_ctc=69.921, loss_att=53.940, acc=0.737, loss=58.717, grad_norm=5.473, loss_scale=1.000, learning_rate=6.221e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 20:21:22,006 (deepspeed_trainer:228) INFO: 42epoch:train:5101-5200batch: iter_time=1.128e-04, loss_ctc=65.235, loss_att=49.051, acc=0.737, loss=53.888, grad_norm=5.752, loss_scale=1.000, learning_rate=6.221e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:21:58,891 (deepspeed_trainer:228) INFO: 42epoch:train:5201-5300batch: iter_time=1.119e-04, loss_ctc=72.060, loss_att=58.276, acc=0.726, loss=62.415, grad_norm=6.041, loss_scale=1.000, learning_rate=6.220e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:22:35,433 (deepspeed_trainer:228) INFO: 42epoch:train:5301-5400batch: iter_time=1.112e-04, loss_ctc=64.098, loss_att=45.582, acc=0.736, loss=51.123, grad_norm=5.442, loss_scale=1.000, learning_rate=6.220e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 20:23:12,467 (deepspeed_trainer:228) INFO: 42epoch:train:5401-5500batch: iter_time=1.140e-04, loss_ctc=71.106, loss_att=51.439, acc=0.738, loss=57.343, grad_norm=7.102, loss_scale=1.000, learning_rate=6.219e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 20:23:49,572 (deepspeed_trainer:228) INFO: 42epoch:train:5501-5600batch: iter_time=1.109e-04, loss_ctc=66.514, loss_att=47.711, acc=0.723, loss=53.329, grad_norm=5.703, loss_scale=1.000, learning_rate=6.219e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 20:24:02,057 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-018:0/16] 2024-12-10 20:24:28,934 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 20:24:45,599 (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: ) [cnode7-018:0/16] 2024-12-10 20:24:45,599 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 20:24:45,601 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 20:25:12,152] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:12,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:12,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:12,153] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:13,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:14,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:14,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:14,126] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:14,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:25:15,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:00,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:00,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:01,060] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:01,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:01,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:01,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:02,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:02,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:02,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:02,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:03,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:02,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:03,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:03,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:03,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:05,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:48,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:48,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:50,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:50,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:50,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:50,567] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:50,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:51,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:51,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:51,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:51,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:51,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:52,113] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:52,424] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:52,462] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:26:55,360] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:36,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:37,241] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:37,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:37,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:38,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:38,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:39,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:39,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:40,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:27:44,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 20:28:38,563 (deepspeed_trainer:228) INFO: 42epoch:train:5601-5700batch: iter_time=2.469, loss_ctc=72.115, loss_att=54.848, acc=0.725, loss=60.015, grad_norm=6.843, loss_scale=1.000, learning_rate=6.218e-05, step_time=0.421 [cnode7-018:0/16] 2024-12-10 20:29:15,186 (deepspeed_trainer:228) INFO: 42epoch:train:5701-5800batch: iter_time=1.123e-04, loss_ctc=67.129, loss_att=50.696, acc=0.718, loss=55.626, grad_norm=6.821, loss_scale=1.000, learning_rate=6.218e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:29:51,693 (deepspeed_trainer:228) INFO: 42epoch:train:5801-5900batch: iter_time=1.133e-04, loss_ctc=65.644, loss_att=47.783, acc=0.720, loss=53.133, grad_norm=6.320, loss_scale=1.000, learning_rate=6.217e-05, step_time=0.365 [2024-12-10 20:30:28,534] [INFO] [logging.py:129:log_dist] [Rank 0] step=621000, skipped=0, lr=[np.float64(6.216693716214642e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:30:28,535] [INFO] [timer.py:264:stop] epoch=0/micro_step=21000/global_step=21000, RunningAvgSamplesPerSec=44.30798083650964, CurrSamplesPerSec=46.60052368003334, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:30:28,536 (deepspeed_trainer:228) INFO: 42epoch:train:5901-6000batch: iter_time=1.149e-04, loss_ctc=67.692, loss_att=51.128, acc=0.733, loss=56.116, grad_norm=6.512, loss_scale=1.000, learning_rate=6.217e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:31:05,180 (deepspeed_trainer:228) INFO: 42epoch:train:6001-6100batch: iter_time=1.193e-04, loss_ctc=68.955, loss_att=52.212, acc=0.716, loss=57.230, grad_norm=6.568, loss_scale=1.000, learning_rate=6.216e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:31:42,109 (deepspeed_trainer:228) INFO: 42epoch:train:6101-6200batch: iter_time=1.107e-04, loss_ctc=64.212, loss_att=46.301, acc=0.741, loss=51.667, grad_norm=5.674, loss_scale=1.000, learning_rate=6.216e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:32:19,322 (deepspeed_trainer:228) INFO: 42epoch:train:6201-6300batch: iter_time=1.084e-04, loss_ctc=72.626, loss_att=54.724, acc=0.731, loss=60.125, grad_norm=6.679, loss_scale=1.000, learning_rate=6.215e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 20:32:56,483 (deepspeed_trainer:228) INFO: 42epoch:train:6301-6400batch: iter_time=1.073e-04, loss_ctc=68.481, loss_att=52.126, acc=0.737, loss=57.044, grad_norm=6.448, loss_scale=1.000, learning_rate=6.215e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 20:33:32,868 (deepspeed_trainer:228) INFO: 42epoch:train:6401-6500batch: iter_time=1.149e-04, loss_ctc=65.854, loss_att=49.482, acc=0.734, loss=54.404, grad_norm=6.218, loss_scale=1.000, learning_rate=6.214e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 20:34:09,218 (deepspeed_trainer:228) INFO: 42epoch:train:6501-6600batch: iter_time=1.116e-04, loss_ctc=66.976, loss_att=50.206, acc=0.713, loss=55.234, grad_norm=5.965, loss_scale=1.000, learning_rate=6.214e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 20:34:45,691 (deepspeed_trainer:228) INFO: 42epoch:train:6601-6700batch: iter_time=1.103e-04, loss_ctc=68.066, loss_att=54.455, acc=0.719, loss=58.551, grad_norm=5.961, loss_scale=1.000, learning_rate=6.213e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 20:35:22,021 (deepspeed_trainer:228) INFO: 42epoch:train:6701-6800batch: iter_time=1.155e-04, loss_ctc=64.064, loss_att=42.777, acc=0.745, loss=49.158, grad_norm=5.806, loss_scale=1.000, learning_rate=6.213e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 20:35:58,505 (deepspeed_trainer:228) INFO: 42epoch:train:6801-6900batch: iter_time=1.140e-04, loss_ctc=67.367, loss_att=52.472, acc=0.732, loss=56.913, grad_norm=5.735, loss_scale=1.000, learning_rate=6.212e-05, step_time=0.364 [2024-12-10 20:36:35,077] [INFO] [logging.py:129:log_dist] [Rank 0] step=622000, skipped=0, lr=[np.float64(6.211694371845503e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:36:35,077] [INFO] [timer.py:264:stop] epoch=0/micro_step=22000/global_step=22000, RunningAvgSamplesPerSec=44.33031303161145, CurrSamplesPerSec=42.003703999033796, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:36:35,079 (deepspeed_trainer:228) INFO: 42epoch:train:6901-7000batch: iter_time=1.166e-04, loss_ctc=71.792, loss_att=52.587, acc=0.731, loss=58.357, grad_norm=6.272, loss_scale=1.000, learning_rate=6.212e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 20:37:11,419 (deepspeed_trainer:228) INFO: 42epoch:train:7001-7100batch: iter_time=1.163e-04, loss_ctc=65.558, loss_att=47.239, acc=0.743, loss=52.733, grad_norm=5.644, loss_scale=1.000, learning_rate=6.211e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 20:37:48,058 (deepspeed_trainer:228) INFO: 42epoch:train:7101-7200batch: iter_time=1.149e-04, loss_ctc=67.604, loss_att=54.525, acc=0.727, loss=58.441, grad_norm=5.809, loss_scale=1.000, learning_rate=6.211e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:38:24,328 (deepspeed_trainer:228) INFO: 42epoch:train:7201-7300batch: iter_time=1.140e-04, loss_ctc=62.418, loss_att=43.064, acc=0.739, loss=48.879, grad_norm=5.224, loss_scale=1.000, learning_rate=6.210e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 20:39:01,500 (deepspeed_trainer:228) INFO: 42epoch:train:7301-7400batch: iter_time=1.111e-04, loss_ctc=77.337, loss_att=56.240, acc=0.724, loss=62.546, grad_norm=5.951, loss_scale=1.000, learning_rate=6.210e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 20:39:37,789 (deepspeed_trainer:228) INFO: 42epoch:train:7401-7500batch: iter_time=1.169e-04, loss_ctc=67.534, loss_att=47.443, acc=0.726, loss=53.469, grad_norm=5.606, loss_scale=1.000, learning_rate=6.209e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 20:39:41,147 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-018:0/16] 2024-12-10 20:40:08,404 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 20:40:25,151 (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: ) [cnode7-018:0/16] 2024-12-10 20:40:25,151 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 20:40:25,154 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 20:40:51,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:51,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:51,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:51,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:51,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:53,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:52,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:53,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:53,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:53,959] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:40:54,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:39,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:40,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:40,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:39,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:40,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:40,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:40,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,461] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:41,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:42,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:41:43,289] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:28,327] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:28,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,333] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:28,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,127] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:30,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,354] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:30,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:29,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:30,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:30,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:31,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:32,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:42:33,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:17,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:17,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:16,501] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:17,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:17,197] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:17,786] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,685] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,070] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,679] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:18,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:19,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:19,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:23,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:43:23,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 20:44:25,669 (deepspeed_trainer:228) INFO: 42epoch:train:7501-7600batch: iter_time=2.455, loss_ctc=70.931, loss_att=52.252, acc=0.727, loss=57.851, grad_norm=6.407, loss_scale=1.000, learning_rate=6.209e-05, step_time=0.424 [cnode7-018:0/16] 2024-12-10 20:45:02,285 (deepspeed_trainer:228) INFO: 42epoch:train:7601-7700batch: iter_time=1.046e-04, loss_ctc=62.007, loss_att=46.684, acc=0.724, loss=51.287, grad_norm=5.904, loss_scale=1.000, learning_rate=6.208e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:45:38,618 (deepspeed_trainer:228) INFO: 42epoch:train:7701-7800batch: iter_time=1.078e-04, loss_ctc=65.003, loss_att=47.042, acc=0.728, loss=52.406, grad_norm=6.050, loss_scale=1.000, learning_rate=6.208e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 20:46:15,316 (deepspeed_trainer:228) INFO: 42epoch:train:7801-7900batch: iter_time=1.124e-04, loss_ctc=67.302, loss_att=50.736, acc=0.727, loss=55.734, grad_norm=6.237, loss_scale=1.000, learning_rate=6.207e-05, step_time=0.367 [2024-12-10 20:46:52,110] [INFO] [logging.py:129:log_dist] [Rank 0] step=623000, skipped=0, lr=[np.float64(6.206707069237149e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:46:52,111] [INFO] [timer.py:264:stop] epoch=0/micro_step=23000/global_step=23000, RunningAvgSamplesPerSec=44.32334307935201, CurrSamplesPerSec=45.95467663805608, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:46:52,112 (deepspeed_trainer:228) INFO: 42epoch:train:7901-8000batch: iter_time=1.118e-04, loss_ctc=68.825, loss_att=50.597, acc=0.727, loss=56.078, grad_norm=6.161, loss_scale=1.000, learning_rate=6.207e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:47:28,857 (deepspeed_trainer:228) INFO: 42epoch:train:8001-8100batch: iter_time=1.109e-04, loss_ctc=71.846, loss_att=52.217, acc=0.733, loss=58.113, grad_norm=5.851, loss_scale=1.000, learning_rate=6.206e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 20:48:05,780 (deepspeed_trainer:228) INFO: 42epoch:train:8101-8200batch: iter_time=1.106e-04, loss_ctc=66.549, loss_att=51.675, acc=0.734, loss=56.145, grad_norm=6.651, loss_scale=1.000, learning_rate=6.206e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:48:42,602 (deepspeed_trainer:228) INFO: 42epoch:train:8201-8300batch: iter_time=1.117e-04, loss_ctc=64.678, loss_att=48.588, acc=0.742, loss=53.415, grad_norm=5.551, loss_scale=1.000, learning_rate=6.205e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:49:19,415 (deepspeed_trainer:228) INFO: 42epoch:train:8301-8400batch: iter_time=1.116e-04, loss_ctc=67.504, loss_att=50.682, acc=0.732, loss=55.736, grad_norm=5.291, loss_scale=1.000, learning_rate=6.205e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:49:56,170 (deepspeed_trainer:228) INFO: 42epoch:train:8401-8500batch: iter_time=1.114e-04, loss_ctc=69.384, loss_att=52.663, acc=0.709, loss=57.661, grad_norm=6.051, loss_scale=1.000, learning_rate=6.204e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 20:50:32,718 (deepspeed_trainer:228) INFO: 42epoch:train:8501-8600batch: iter_time=1.107e-04, loss_ctc=67.563, loss_att=52.281, acc=0.725, loss=56.860, grad_norm=6.191, loss_scale=1.000, learning_rate=6.204e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:51:09,390 (deepspeed_trainer:228) INFO: 42epoch:train:8601-8700batch: iter_time=1.125e-04, loss_ctc=63.803, loss_att=43.870, acc=0.747, loss=49.829, grad_norm=5.127, loss_scale=1.000, learning_rate=6.203e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:51:46,277 (deepspeed_trainer:228) INFO: 42epoch:train:8701-8800batch: iter_time=1.138e-04, loss_ctc=68.057, loss_att=51.159, acc=0.737, loss=56.218, grad_norm=5.784, loss_scale=1.000, learning_rate=6.203e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:52:23,077 (deepspeed_trainer:228) INFO: 42epoch:train:8801-8900batch: iter_time=1.105e-04, loss_ctc=69.119, loss_att=54.104, acc=0.726, loss=58.587, grad_norm=5.559, loss_scale=1.000, learning_rate=6.202e-05, step_time=0.368 [2024-12-10 20:52:59,767] [INFO] [logging.py:129:log_dist] [Rank 0] step=624000, skipped=0, lr=[np.float64(6.201731760126068e-05)], mom=[[0.9, 0.98]] [2024-12-10 20:52:59,768] [INFO] [timer.py:264:stop] epoch=0/micro_step=24000/global_step=24000, RunningAvgSamplesPerSec=44.33706292525834, CurrSamplesPerSec=44.9789144825842, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 20:52:59,770 (deepspeed_trainer:228) INFO: 42epoch:train:8901-9000batch: iter_time=1.105e-04, loss_ctc=64.194, loss_att=44.049, acc=0.751, loss=50.093, grad_norm=5.913, loss_scale=1.000, learning_rate=6.202e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 20:53:36,700 (deepspeed_trainer:228) INFO: 42epoch:train:9001-9100batch: iter_time=1.121e-04, loss_ctc=71.105, loss_att=58.367, acc=0.718, loss=62.181, grad_norm=6.488, loss_scale=1.000, learning_rate=6.201e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 20:54:13,267 (deepspeed_trainer:228) INFO: 42epoch:train:9101-9200batch: iter_time=1.127e-04, loss_ctc=58.705, loss_att=40.931, acc=0.741, loss=46.277, grad_norm=5.141, loss_scale=1.000, learning_rate=6.201e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 20:54:50,143 (deepspeed_trainer:228) INFO: 42epoch:train:9201-9300batch: iter_time=1.092e-04, loss_ctc=79.168, loss_att=55.605, acc=0.723, loss=62.677, grad_norm=5.777, loss_scale=1.000, learning_rate=6.200e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 20:55:21,152 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-018:0/16] 2024-12-10 20:55:48,354 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 20:56:04,891 (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: ) [cnode7-018:0/16] 2024-12-10 20:56:04,891 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 20:56:04,893 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 20:56:32,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:32,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:32,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:32,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:32,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:34,263] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:34,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:34,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:34,464] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:33,716] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:56:34,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:21,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:21,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:22,229] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:22,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:21,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:21,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:22,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,424] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,482] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:23,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:57:32,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:09,610] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:10,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:11,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:10,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:10,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:11,519] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:10,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:11,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:12,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:11,590] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:13,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:12,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:12,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:12,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:13,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:27,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:58,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:57,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:59,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:58,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:59,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:59,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:00,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:58:59,586] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:01,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:00,334] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:01,663] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:00,855] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:01,539] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:01,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:01,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 20:59:19,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 20:59:54,445 (deepspeed_trainer:228) INFO: 42epoch:train:9301-9400batch: iter_time=2.500, loss_ctc=66.500, loss_att=49.228, acc=0.728, loss=54.375, grad_norm=6.779, loss_scale=1.000, learning_rate=6.200e-05, step_time=0.543 [cnode7-018:0/16] 2024-12-10 21:00:31,541 (deepspeed_trainer:228) INFO: 42epoch:train:9401-9500batch: iter_time=1.102e-04, loss_ctc=69.746, loss_att=50.930, acc=0.730, loss=56.565, grad_norm=6.123, loss_scale=1.000, learning_rate=6.199e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:01:08,376 (deepspeed_trainer:228) INFO: 42epoch:train:9501-9600batch: iter_time=1.100e-04, loss_ctc=64.859, loss_att=47.331, acc=0.724, loss=52.587, grad_norm=6.145, loss_scale=1.000, learning_rate=6.199e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:01:45,399 (deepspeed_trainer:228) INFO: 42epoch:train:9601-9700batch: iter_time=1.080e-04, loss_ctc=65.508, loss_att=49.032, acc=0.730, loss=53.981, grad_norm=6.033, loss_scale=1.000, learning_rate=6.199e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:02:22,636 (deepspeed_trainer:228) INFO: 42epoch:train:9701-9800batch: iter_time=1.081e-04, loss_ctc=66.566, loss_att=49.996, acc=0.725, loss=54.948, grad_norm=5.960, loss_scale=1.000, learning_rate=6.198e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 21:02:59,638 (deepspeed_trainer:228) INFO: 42epoch:train:9801-9900batch: iter_time=1.103e-04, loss_ctc=66.682, loss_att=48.197, acc=0.728, loss=53.737, grad_norm=5.748, loss_scale=1.000, learning_rate=6.198e-05, step_time=0.369 [2024-12-10 21:03:36,804] [INFO] [logging.py:129:log_dist] [Rank 0] step=625000, skipped=0, lr=[np.float64(6.196768396519133e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:03:36,804] [INFO] [timer.py:264:stop] epoch=0/micro_step=25000/global_step=25000, RunningAvgSamplesPerSec=44.25425475948358, CurrSamplesPerSec=41.83199711258655, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:03:36,806 (deepspeed_trainer:228) INFO: 42epoch:train:9901-10000batch: iter_time=1.094e-04, loss_ctc=74.626, loss_att=55.305, acc=0.733, loss=61.090, grad_norm=5.849, loss_scale=1.000, learning_rate=6.197e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 21:04:14,107 (deepspeed_trainer:228) INFO: 42epoch:train:10001-10100batch: iter_time=1.110e-04, loss_ctc=63.372, loss_att=48.720, acc=0.738, loss=53.112, grad_norm=6.979, loss_scale=1.000, learning_rate=6.197e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 21:04:51,024 (deepspeed_trainer:228) INFO: 42epoch:train:10101-10200batch: iter_time=1.107e-04, loss_ctc=67.936, loss_att=49.943, acc=0.741, loss=55.357, grad_norm=5.665, loss_scale=1.000, learning_rate=6.196e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:05:28,076 (deepspeed_trainer:228) INFO: 42epoch:train:10201-10300batch: iter_time=1.093e-04, loss_ctc=68.471, loss_att=53.015, acc=0.722, loss=57.650, grad_norm=5.805, loss_scale=1.000, learning_rate=6.196e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:06:05,020 (deepspeed_trainer:228) INFO: 42epoch:train:10301-10400batch: iter_time=1.097e-04, loss_ctc=67.900, loss_att=50.483, acc=0.724, loss=55.708, grad_norm=5.889, loss_scale=1.000, learning_rate=6.195e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 21:06:41,860 (deepspeed_trainer:228) INFO: 42epoch:train:10401-10500batch: iter_time=1.135e-04, loss_ctc=61.380, loss_att=45.028, acc=0.734, loss=49.924, grad_norm=5.763, loss_scale=1.000, learning_rate=6.195e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:07:19,333 (deepspeed_trainer:228) INFO: 42epoch:train:10501-10600batch: iter_time=1.074e-04, loss_ctc=66.616, loss_att=47.614, acc=0.740, loss=53.332, grad_norm=5.598, loss_scale=1.000, learning_rate=6.194e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 21:07:57,084 (deepspeed_trainer:228) INFO: 42epoch:train:10601-10700batch: iter_time=1.104e-04, loss_ctc=67.843, loss_att=52.651, acc=0.735, loss=57.230, grad_norm=5.854, loss_scale=1.000, learning_rate=6.194e-05, step_time=0.377 [cnode7-018:0/16] 2024-12-10 21:08:34,836 (deepspeed_trainer:228) INFO: 42epoch:train:10701-10800batch: iter_time=1.093e-04, loss_ctc=68.589, loss_att=51.313, acc=0.737, loss=56.479, grad_norm=5.523, loss_scale=1.000, learning_rate=6.193e-05, step_time=0.377 [cnode7-018:0/16] 2024-12-10 21:09:11,959 (deepspeed_trainer:228) INFO: 42epoch:train:10801-10900batch: iter_time=1.099e-04, loss_ctc=70.434, loss_att=53.935, acc=0.736, loss=58.897, grad_norm=5.919, loss_scale=1.000, learning_rate=6.193e-05, step_time=0.371 [2024-12-10 21:09:48,605] [INFO] [logging.py:129:log_dist] [Rank 0] step=626000, skipped=0, lr=[np.float64(6.191816930691656e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:09:48,605] [INFO] [timer.py:264:stop] epoch=0/micro_step=26000/global_step=26000, RunningAvgSamplesPerSec=44.25078200436428, CurrSamplesPerSec=47.01932923547708, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:09:48,607 (deepspeed_trainer:228) INFO: 42epoch:train:10901-11000batch: iter_time=1.088e-04, loss_ctc=64.377, loss_att=48.420, acc=0.726, loss=53.209, grad_norm=5.441, loss_scale=1.000, learning_rate=6.192e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 21:10:25,385 (deepspeed_trainer:228) INFO: 42epoch:train:11001-11100batch: iter_time=1.094e-04, loss_ctc=61.973, loss_att=43.962, acc=0.741, loss=49.363, grad_norm=5.071, loss_scale=1.000, learning_rate=6.192e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 21:11:02,240 (deepspeed_trainer:228) INFO: 42epoch:train:11101-11200batch: iter_time=1.084e-04, loss_ctc=72.694, loss_att=49.463, acc=0.729, loss=56.439, grad_norm=5.576, loss_scale=1.000, learning_rate=6.191e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 21:11:24,106 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-018:0/16] 2024-12-10 21:11:51,094 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 21:12:07,163 (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: ) [cnode7-018:0/16] 2024-12-10 21:12:07,163 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 21:12:07,165 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 21:12:34,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:33,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:33,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:34,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:36,006] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:36,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:36,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:35,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:37,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:36,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:12:36,153] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:22,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:21,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:23,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:22,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:23,800] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:22,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:24,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:24,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:23,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:23,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:25,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:24,264] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:25,486] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:25,586] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:25,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:13:25,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:10,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:10,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:11,289] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:12,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:12,596] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:11,743] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:12,859] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:11,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:13,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:13,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:13,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:12,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:12,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:14,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:14,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:17,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:58,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:58,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:00,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:14:59,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:00,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:01,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:00,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:00,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:01,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:00,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:02,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:02,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:01,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:02,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:01,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:15:09,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 21:15:52,977 (deepspeed_trainer:228) INFO: 42epoch:train:11201-11300batch: iter_time=2.458, loss_ctc=71.380, loss_att=56.715, acc=0.721, loss=61.119, grad_norm=6.147, loss_scale=1.000, learning_rate=6.191e-05, step_time=0.450 [cnode7-018:0/16] 2024-12-10 21:16:30,096 (deepspeed_trainer:228) INFO: 42epoch:train:11301-11400batch: iter_time=1.040e-04, loss_ctc=67.851, loss_att=51.329, acc=0.726, loss=56.269, grad_norm=5.932, loss_scale=1.000, learning_rate=6.190e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 21:17:06,820 (deepspeed_trainer:228) INFO: 42epoch:train:11401-11500batch: iter_time=1.043e-04, loss_ctc=65.106, loss_att=47.979, acc=0.723, loss=53.138, grad_norm=5.753, loss_scale=1.000, learning_rate=6.190e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 21:17:43,898 (deepspeed_trainer:228) INFO: 42epoch:train:11501-11600batch: iter_time=1.087e-04, loss_ctc=66.968, loss_att=49.811, acc=0.747, loss=54.967, grad_norm=5.583, loss_scale=1.000, learning_rate=6.189e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:18:20,920 (deepspeed_trainer:228) INFO: 42epoch:train:11601-11700batch: iter_time=1.090e-04, loss_ctc=67.230, loss_att=51.491, acc=0.723, loss=56.204, grad_norm=5.955, loss_scale=1.000, learning_rate=6.189e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:18:57,911 (deepspeed_trainer:228) INFO: 42epoch:train:11701-11800batch: iter_time=1.119e-04, loss_ctc=63.317, loss_att=46.664, acc=0.741, loss=51.653, grad_norm=5.457, loss_scale=1.000, learning_rate=6.188e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:19:35,001 (deepspeed_trainer:228) INFO: 42epoch:train:11801-11900batch: iter_time=1.085e-04, loss_ctc=74.761, loss_att=56.417, acc=0.739, loss=61.912, grad_norm=6.483, loss_scale=1.000, learning_rate=6.188e-05, step_time=0.370 [2024-12-10 21:20:12,160] [INFO] [logging.py:129:log_dist] [Rank 0] step=627000, skipped=0, lr=[np.float64(6.186877315185459e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:20:12,161] [INFO] [timer.py:264:stop] epoch=0/micro_step=27000/global_step=27000, RunningAvgSamplesPerSec=44.21960547430332, CurrSamplesPerSec=44.52583939270477, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:20:12,162 (deepspeed_trainer:228) INFO: 42epoch:train:11901-12000batch: iter_time=1.124e-04, loss_ctc=66.824, loss_att=52.882, acc=0.739, loss=57.026, grad_norm=6.293, loss_scale=1.000, learning_rate=6.187e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 21:20:49,117 (deepspeed_trainer:228) INFO: 42epoch:train:12001-12100batch: iter_time=1.105e-04, loss_ctc=66.254, loss_att=50.759, acc=0.744, loss=55.414, grad_norm=5.733, loss_scale=1.000, learning_rate=6.187e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:21:26,037 (deepspeed_trainer:228) INFO: 42epoch:train:12101-12200batch: iter_time=1.087e-04, loss_ctc=66.812, loss_att=51.380, acc=0.721, loss=56.009, grad_norm=6.144, loss_scale=1.000, learning_rate=6.186e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:22:03,060 (deepspeed_trainer:228) INFO: 42epoch:train:12201-12300batch: iter_time=1.089e-04, loss_ctc=67.828, loss_att=53.447, acc=0.733, loss=57.765, grad_norm=6.368, loss_scale=1.000, learning_rate=6.186e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:22:39,639 (deepspeed_trainer:228) INFO: 42epoch:train:12301-12400batch: iter_time=1.079e-04, loss_ctc=61.313, loss_att=44.523, acc=0.743, loss=49.567, grad_norm=5.699, loss_scale=1.000, learning_rate=6.185e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 21:23:16,645 (deepspeed_trainer:228) INFO: 42epoch:train:12401-12500batch: iter_time=1.096e-04, loss_ctc=68.254, loss_att=48.125, acc=0.748, loss=54.157, grad_norm=5.546, loss_scale=1.000, learning_rate=6.185e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:23:53,456 (deepspeed_trainer:228) INFO: 42epoch:train:12501-12600batch: iter_time=1.128e-04, loss_ctc=70.137, loss_att=54.917, acc=0.736, loss=59.482, grad_norm=5.659, loss_scale=1.000, learning_rate=6.184e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 21:24:29,990 (deepspeed_trainer:228) INFO: 42epoch:train:12601-12700batch: iter_time=1.089e-04, loss_ctc=64.521, loss_att=48.705, acc=0.742, loss=53.448, grad_norm=5.222, loss_scale=1.000, learning_rate=6.184e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 21:25:06,874 (deepspeed_trainer:228) INFO: 42epoch:train:12701-12800batch: iter_time=1.090e-04, loss_ctc=71.433, loss_att=58.041, acc=0.729, loss=62.040, grad_norm=5.642, loss_scale=1.000, learning_rate=6.183e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:25:43,977 (deepspeed_trainer:228) INFO: 42epoch:train:12801-12900batch: iter_time=1.169e-04, loss_ctc=64.204, loss_att=45.574, acc=0.736, loss=51.162, grad_norm=5.241, loss_scale=1.000, learning_rate=6.183e-05, step_time=0.369 [2024-12-10 21:26:20,947] [INFO] [logging.py:129:log_dist] [Rank 0] step=628000, skipped=0, lr=[np.float64(6.181949502806968e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:26:20,947] [INFO] [timer.py:264:stop] epoch=0/micro_step=28000/global_step=28000, RunningAvgSamplesPerSec=44.228448910485575, CurrSamplesPerSec=44.11741666971886, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:26:20,949 (deepspeed_trainer:228) INFO: 42epoch:train:12901-13000batch: iter_time=1.083e-04, loss_ctc=70.561, loss_att=51.091, acc=0.740, loss=56.932, grad_norm=5.296, loss_scale=1.000, learning_rate=6.182e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:26:57,640 (deepspeed_trainer:228) INFO: 42epoch:train:13001-13100batch: iter_time=1.098e-04, loss_ctc=65.989, loss_att=47.777, acc=0.725, loss=53.242, grad_norm=5.457, loss_scale=1.000, learning_rate=6.182e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 21:27:10,272 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-018:0/16] 2024-12-10 21:27:37,148 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 21:27:52,586 (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: ) [cnode7-018:0/16] 2024-12-10 21:27:52,586 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 21:27:52,588 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 21:28:19,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:19,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:20,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:20,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:21,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:21,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:21,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,246] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:22,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:28:23,227] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:07,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:08,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:08,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:09,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:09,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:09,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:10,271] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:10,362] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:11,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:11,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:11,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:11,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:12,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:12,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:12,525] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:12,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:56,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:56,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:57,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:58,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:58,688] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:59,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:59,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:59,424] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:29:59,798] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:00,121] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:00,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:01,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:01,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:01,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:02,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:02,256] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:43,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:45,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:46,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:46,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:46,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:47,321] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:47,707] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:48,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:48,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:49,398] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:49,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:49,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:50,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:50,332] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:51,016] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:30:52,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 21:31:46,163 (deepspeed_trainer:228) INFO: 42epoch:train:13101-13200batch: iter_time=2.473, loss_ctc=72.503, loss_att=54.329, acc=0.731, loss=59.799, grad_norm=6.264, loss_scale=1.000, learning_rate=6.181e-05, step_time=0.412 [cnode7-018:0/16] 2024-12-10 21:32:23,087 (deepspeed_trainer:228) INFO: 42epoch:train:13201-13300batch: iter_time=1.085e-04, loss_ctc=66.296, loss_att=50.120, acc=0.728, loss=54.956, grad_norm=5.595, loss_scale=1.000, learning_rate=6.181e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:32:59,710 (deepspeed_trainer:228) INFO: 42epoch:train:13301-13400batch: iter_time=1.060e-04, loss_ctc=64.925, loss_att=47.635, acc=0.728, loss=52.821, grad_norm=6.022, loss_scale=1.000, learning_rate=6.180e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 21:33:36,469 (deepspeed_trainer:228) INFO: 42epoch:train:13401-13500batch: iter_time=1.126e-04, loss_ctc=67.733, loss_att=52.034, acc=0.743, loss=56.753, grad_norm=5.627, loss_scale=1.000, learning_rate=6.180e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 21:34:13,445 (deepspeed_trainer:228) INFO: 42epoch:train:13501-13600batch: iter_time=1.108e-04, loss_ctc=68.414, loss_att=51.360, acc=0.722, loss=56.479, grad_norm=5.962, loss_scale=1.000, learning_rate=6.179e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:34:50,307 (deepspeed_trainer:228) INFO: 42epoch:train:13601-13700batch: iter_time=1.112e-04, loss_ctc=64.018, loss_att=46.625, acc=0.749, loss=51.858, grad_norm=5.442, loss_scale=1.000, learning_rate=6.179e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 21:35:27,894 (deepspeed_trainer:228) INFO: 42epoch:train:13701-13800batch: iter_time=1.159e-04, loss_ctc=71.844, loss_att=53.968, acc=0.741, loss=59.337, grad_norm=6.251, loss_scale=1.000, learning_rate=6.178e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 21:36:05,148 (deepspeed_trainer:228) INFO: 42epoch:train:13801-13900batch: iter_time=1.207e-04, loss_ctc=67.703, loss_att=52.441, acc=0.741, loss=57.027, grad_norm=5.641, loss_scale=1.000, learning_rate=6.178e-05, step_time=0.372 [2024-12-10 21:36:42,733] [INFO] [logging.py:129:log_dist] [Rank 0] step=629000, skipped=0, lr=[np.float64(6.177033446625316e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:36:42,734] [INFO] [timer.py:264:stop] epoch=0/micro_step=29000/global_step=29000, RunningAvgSamplesPerSec=44.21331437128659, CurrSamplesPerSec=42.573603108908465, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:36:42,735 (deepspeed_trainer:228) INFO: 42epoch:train:13901-14000batch: iter_time=1.123e-04, loss_ctc=65.424, loss_att=51.128, acc=0.740, loss=55.430, grad_norm=5.481, loss_scale=1.000, learning_rate=6.177e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 21:37:20,043 (deepspeed_trainer:228) INFO: 42epoch:train:14001-14100batch: iter_time=1.093e-04, loss_ctc=66.590, loss_att=51.603, acc=0.717, loss=56.088, grad_norm=5.759, loss_scale=1.000, learning_rate=6.177e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 21:37:57,469 (deepspeed_trainer:228) INFO: 42epoch:train:14101-14200batch: iter_time=1.092e-04, loss_ctc=67.762, loss_att=53.772, acc=0.730, loss=57.972, grad_norm=5.586, loss_scale=1.000, learning_rate=6.176e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 21:38:35,058 (deepspeed_trainer:228) INFO: 42epoch:train:14201-14300batch: iter_time=1.113e-04, loss_ctc=63.827, loss_att=42.323, acc=0.754, loss=48.765, grad_norm=5.991, loss_scale=1.000, learning_rate=6.176e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 21:39:12,343 (deepspeed_trainer:228) INFO: 42epoch:train:14301-14400batch: iter_time=1.070e-04, loss_ctc=67.089, loss_att=52.276, acc=0.740, loss=56.747, grad_norm=5.281, loss_scale=1.000, learning_rate=6.175e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 21:39:49,288 (deepspeed_trainer:228) INFO: 42epoch:train:14401-14500batch: iter_time=1.078e-04, loss_ctc=71.075, loss_att=54.000, acc=0.735, loss=59.113, grad_norm=5.399, loss_scale=1.000, learning_rate=6.175e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:40:26,322 (deepspeed_trainer:228) INFO: 42epoch:train:14501-14600batch: iter_time=1.107e-04, loss_ctc=64.931, loss_att=46.650, acc=0.749, loss=52.136, grad_norm=5.355, loss_scale=1.000, learning_rate=6.174e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:41:03,853 (deepspeed_trainer:228) INFO: 42epoch:train:14601-14700batch: iter_time=1.085e-04, loss_ctc=67.144, loss_att=54.275, acc=0.730, loss=58.123, grad_norm=6.329, loss_scale=1.000, learning_rate=6.174e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 21:41:40,995 (deepspeed_trainer:228) INFO: 42epoch:train:14701-14800batch: iter_time=1.082e-04, loss_ctc=61.904, loss_att=43.192, acc=0.745, loss=48.780, grad_norm=5.156, loss_scale=1.000, learning_rate=6.173e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 21:42:18,078 (deepspeed_trainer:228) INFO: 42epoch:train:14801-14900batch: iter_time=1.088e-04, loss_ctc=76.810, loss_att=56.974, acc=0.725, loss=62.944, grad_norm=5.870, loss_scale=1.000, learning_rate=6.173e-05, step_time=0.371 [2024-12-10 21:42:55,918] [INFO] [logging.py:129:log_dist] [Rank 0] step=630000, skipped=0, lr=[np.float64(6.172129099970461e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:42:55,919] [INFO] [timer.py:264:stop] epoch=0/micro_step=30000/global_step=30000, RunningAvgSamplesPerSec=44.2051166773961, CurrSamplesPerSec=46.17897394861437, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:42:55,920 (deepspeed_trainer:228) INFO: 42epoch:train:14901-15000batch: iter_time=1.086e-04, loss_ctc=66.862, loss_att=47.391, acc=0.730, loss=53.224, grad_norm=5.262, loss_scale=1.000, learning_rate=6.172e-05, step_time=0.378 [2024-12-10 21:43:09,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:10,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:10,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,371] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,301] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,381] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,653] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,595] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:11,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:26,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:26,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:27,329] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:28,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:28,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:28,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:28,710] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,377] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:29,690] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:30,273] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:30,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:30,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:42,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:42,809] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:42,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:43,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:44,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:44,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:44,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:44,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:45,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:45,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:45,683] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:46,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:46,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:46,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:46,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:47,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:58,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:58,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:58,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:43:59,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:00,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:00,989] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:00,901] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,245] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,938] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:01,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:02,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:03,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:03,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:44:16,020] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 42 is about to be saved! [2024-12-10 21:44:16,052] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/mp_rank_00_model_states.pt [2024-12-10 21:44:16,052] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/mp_rank_00_model_states.pt... [2024-12-10 21:44:17,961] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/mp_rank_00_model_states.pt. [2024-12-10 21:44:18,006] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,006] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,117] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,119] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,009] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,120] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,120] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,121] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,010] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,122] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,123] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,012] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,014] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,125] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,015] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,015] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-10 21:44:18,909] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,909] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,909] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,915] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,919] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,919] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,919] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,817] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,817] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,817] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,817] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,818] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,818] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,818] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,818] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,818] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,933] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,933] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,934] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,956] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,956] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,956] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,961] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,961] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,961] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,969] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,969] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,969] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:18,972] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-10 21:44:18,972] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-12-10 21:44:18,972] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,286] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,286] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,213] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-10 21:44:19,213] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,213] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,653] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-10 21:44:19,654] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,654] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,667] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-10 21:44:19,667] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,667] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,708] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-10 21:44:19,708] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,708] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [2024-12-10 21:44:19,719] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-10 21:44:19,719] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_42/42/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-12-10 21:44:19,719] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 42 is ready now! [cnode7-018:0/16] 2024-12-10 21:44:19,832 (deepspeed_trainer:158) INFO: 42epoch results: [train] iter_time=0.132, loss_ctc=67.821, loss_att=50.530, acc=0.732, loss=55.716, grad_norm=5.904, loss_scale=1.000, learning_rate=6.209e-05, step_time=0.373, time=2 hours, 6 minutes and 22.58 seconds, total_count=630042, gpu_max_cached_mem_GB=28.656, [valid] loss_ctc=3.219, cer_ctc=0.064, loss_att=6.469, acc=0.846, cer=0.396, wer=1.000, loss=5.500, time=1 minute and 11.89 seconds, total_count=42, gpu_max_cached_mem_GB=28.656 [cnode7-018:0/16] 2024-12-10 21:44:21,025 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-018:0/16] 2024-12-10 21:44:47,274 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 21:45:04,110 (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: ) [cnode7-018:0/16] 2024-12-10 21:45:04,110 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 21:45:04,112 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 21:45:25,085] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,141] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:25,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:26,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:26,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:26,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:26,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:27,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:27,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:27,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:45:28,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:12,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:12,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:13,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:13,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:13,779] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:14,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:14,377] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:14,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:14,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:14,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:15,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:15,673] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:16,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:17,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:17,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:18,989] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:46:59,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:01,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:01,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:01,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:01,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:02,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:02,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:02,443] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:02,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:02,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:03,220] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:04,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:05,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:07,391] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:08,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:09,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:47,500] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:49,281] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:49,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:49,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:50,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:50,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:50,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:51,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:51,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:51,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:52,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:52,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:54,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:57,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:57,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 21:47:59,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 21:49:02,215 (deepspeed_trainer:228) INFO: 43epoch:train:1-100batch: iter_time=2.377, loss_ctc=68.737, loss_att=50.850, acc=0.735, loss=56.178, grad_norm=6.099, loss_scale=1.000, learning_rate=6.172e-05, step_time=0.435 [cnode7-018:0/16] 2024-12-10 21:49:39,850 (deepspeed_trainer:228) INFO: 43epoch:train:101-200batch: iter_time=1.120e-04, loss_ctc=73.255, loss_att=58.790, acc=0.721, loss=63.137, grad_norm=6.611, loss_scale=1.000, learning_rate=6.171e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-10 21:50:17,275 (deepspeed_trainer:228) INFO: 43epoch:train:201-300batch: iter_time=1.129e-04, loss_ctc=78.434, loss_att=58.234, acc=0.728, loss=64.287, grad_norm=6.231, loss_scale=1.000, learning_rate=6.171e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 21:50:54,199 (deepspeed_trainer:228) INFO: 43epoch:train:301-400batch: iter_time=1.087e-04, loss_ctc=69.963, loss_att=50.905, acc=0.732, loss=56.615, grad_norm=6.046, loss_scale=1.000, learning_rate=6.170e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:51:31,204 (deepspeed_trainer:228) INFO: 43epoch:train:401-500batch: iter_time=1.109e-04, loss_ctc=69.771, loss_att=48.910, acc=0.732, loss=55.146, grad_norm=6.886, loss_scale=1.000, learning_rate=6.170e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:52:07,852 (deepspeed_trainer:228) INFO: 43epoch:train:501-600batch: iter_time=1.115e-04, loss_ctc=60.777, loss_att=42.408, acc=0.734, loss=47.915, grad_norm=5.549, loss_scale=1.000, learning_rate=6.169e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 21:52:44,734 (deepspeed_trainer:228) INFO: 43epoch:train:601-700batch: iter_time=1.078e-04, loss_ctc=65.658, loss_att=49.892, acc=0.736, loss=54.632, grad_norm=6.262, loss_scale=1.000, learning_rate=6.169e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 21:53:21,801 (deepspeed_trainer:228) INFO: 43epoch:train:701-800batch: iter_time=1.086e-04, loss_ctc=79.241, loss_att=55.467, acc=0.734, loss=62.606, grad_norm=6.577, loss_scale=1.000, learning_rate=6.168e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:53:59,052 (deepspeed_trainer:228) INFO: 43epoch:train:801-900batch: iter_time=1.098e-04, loss_ctc=66.727, loss_att=45.073, acc=0.729, loss=51.576, grad_norm=6.543, loss_scale=1.000, learning_rate=6.168e-05, step_time=0.372 [2024-12-10 21:54:36,338] [INFO] [logging.py:129:log_dist] [Rank 0] step=631000, skipped=0, lr=[np.float64(6.16723641643134e-05)], mom=[[0.9, 0.98]] [2024-12-10 21:54:36,339] [INFO] [timer.py:264:stop] epoch=0/micro_step=31000/global_step=31000, RunningAvgSamplesPerSec=44.17995902755104, CurrSamplesPerSec=42.779316977646516, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 21:54:36,340 (deepspeed_trainer:228) INFO: 43epoch:train:901-1000batch: iter_time=1.096e-04, loss_ctc=73.512, loss_att=57.098, acc=0.727, loss=62.041, grad_norm=5.858, loss_scale=1.000, learning_rate=6.167e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 21:55:13,332 (deepspeed_trainer:228) INFO: 43epoch:train:1001-1100batch: iter_time=1.112e-04, loss_ctc=66.963, loss_att=49.903, acc=0.744, loss=55.048, grad_norm=5.717, loss_scale=1.000, learning_rate=6.167e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:55:51,163 (deepspeed_trainer:228) INFO: 43epoch:train:1101-1200batch: iter_time=1.124e-04, loss_ctc=69.567, loss_att=51.243, acc=0.725, loss=56.731, grad_norm=6.297, loss_scale=1.000, learning_rate=6.167e-05, step_time=0.378 [cnode7-018:0/16] 2024-12-10 21:56:28,456 (deepspeed_trainer:228) INFO: 43epoch:train:1201-1300batch: iter_time=1.135e-04, loss_ctc=72.402, loss_att=50.127, acc=0.735, loss=56.812, grad_norm=5.995, loss_scale=1.000, learning_rate=6.166e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 21:57:06,043 (deepspeed_trainer:228) INFO: 43epoch:train:1301-1400batch: iter_time=1.069e-04, loss_ctc=74.629, loss_att=52.334, acc=0.728, loss=59.018, grad_norm=6.350, loss_scale=1.000, learning_rate=6.166e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-10 21:57:43,229 (deepspeed_trainer:228) INFO: 43epoch:train:1401-1500batch: iter_time=1.115e-04, loss_ctc=78.031, loss_att=57.121, acc=0.727, loss=63.375, grad_norm=6.626, loss_scale=1.000, learning_rate=6.165e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 21:58:20,552 (deepspeed_trainer:228) INFO: 43epoch:train:1501-1600batch: iter_time=1.107e-04, loss_ctc=79.439, loss_att=57.646, acc=0.732, loss=64.184, grad_norm=6.296, loss_scale=1.000, learning_rate=6.165e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 21:58:57,576 (deepspeed_trainer:228) INFO: 43epoch:train:1601-1700batch: iter_time=1.093e-04, loss_ctc=61.994, loss_att=47.381, acc=0.738, loss=51.764, grad_norm=5.356, loss_scale=1.000, learning_rate=6.164e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 21:59:34,473 (deepspeed_trainer:228) INFO: 43epoch:train:1701-1800batch: iter_time=1.107e-04, loss_ctc=64.869, loss_att=50.851, acc=0.730, loss=55.066, grad_norm=5.801, loss_scale=1.000, learning_rate=6.164e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:00:05,537 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-018:0/16] 2024-12-10 22:00:32,495 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 22:00:48,111 (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: ) [cnode7-018:0/16] 2024-12-10 22:00:48,112 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 22:00:48,114 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 22:01:16,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:16,368] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:16,847] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:16,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:16,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:17,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:17,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:17,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,114] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,156] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,900] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:19,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:01:18,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:05,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:05,949] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:05,986] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:06,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:06,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:06,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,881] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,588] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:08,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:07,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:08,932] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:08,984] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:53,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:53,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:54,520] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:54,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:54,664] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:55,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:54,997] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:55,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:55,830] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:56,359] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:56,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:57,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:57,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:58,604] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:58,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:02:58,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:41,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:42,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:42,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:43,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:44,031] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:43,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:43,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:44,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:44,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:45,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:45,445] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:46,258] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:47,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:47,234] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:47,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:03:49,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 22:04:25,171 (deepspeed_trainer:228) INFO: 43epoch:train:1801-1900batch: iter_time=2.537, loss_ctc=68.999, loss_att=54.540, acc=0.733, loss=58.909, grad_norm=5.742, loss_scale=1.000, learning_rate=6.163e-05, step_time=0.369 [2024-12-10 22:05:01,938] [INFO] [logging.py:129:log_dist] [Rank 0] step=632000, skipped=0, lr=[np.float64(6.162355349854005e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:05:01,938] [INFO] [timer.py:264:stop] epoch=0/micro_step=32000/global_step=32000, RunningAvgSamplesPerSec=44.17826454424493, CurrSamplesPerSec=44.39308920195939, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:05:01,940 (deepspeed_trainer:228) INFO: 43epoch:train:1901-2000batch: iter_time=1.076e-04, loss_ctc=70.565, loss_att=56.422, acc=0.716, loss=60.695, grad_norm=6.247, loss_scale=1.000, learning_rate=6.163e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:05:38,859 (deepspeed_trainer:228) INFO: 43epoch:train:2001-2100batch: iter_time=1.101e-04, loss_ctc=73.206, loss_att=56.308, acc=0.717, loss=61.405, grad_norm=5.767, loss_scale=1.000, learning_rate=6.162e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:06:15,549 (deepspeed_trainer:228) INFO: 43epoch:train:2101-2200batch: iter_time=1.100e-04, loss_ctc=75.305, loss_att=53.956, acc=0.733, loss=60.357, grad_norm=6.355, loss_scale=1.000, learning_rate=6.162e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 22:06:50,776 (deepspeed_trainer:228) INFO: 43epoch:train:2201-2300batch: iter_time=1.124e-04, loss_ctc=67.493, loss_att=49.781, acc=0.721, loss=55.077, grad_norm=6.195, loss_scale=1.000, learning_rate=6.161e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 22:07:27,600 (deepspeed_trainer:228) INFO: 43epoch:train:2301-2400batch: iter_time=1.129e-04, loss_ctc=68.334, loss_att=48.580, acc=0.726, loss=54.523, grad_norm=6.921, loss_scale=1.000, learning_rate=6.161e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:08:04,442 (deepspeed_trainer:228) INFO: 43epoch:train:2401-2500batch: iter_time=1.092e-04, loss_ctc=65.268, loss_att=48.618, acc=0.723, loss=53.616, grad_norm=6.310, loss_scale=1.000, learning_rate=6.160e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:08:41,362 (deepspeed_trainer:228) INFO: 43epoch:train:2501-2600batch: iter_time=1.128e-04, loss_ctc=64.061, loss_att=45.976, acc=0.733, loss=51.398, grad_norm=5.535, loss_scale=1.000, learning_rate=6.160e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:09:18,484 (deepspeed_trainer:228) INFO: 43epoch:train:2601-2700batch: iter_time=1.115e-04, loss_ctc=74.160, loss_att=54.056, acc=0.737, loss=60.068, grad_norm=7.123, loss_scale=1.000, learning_rate=6.159e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 22:09:55,505 (deepspeed_trainer:228) INFO: 43epoch:train:2701-2800batch: iter_time=1.111e-04, loss_ctc=67.971, loss_att=47.346, acc=0.719, loss=53.551, grad_norm=6.711, loss_scale=1.000, learning_rate=6.159e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:10:32,727 (deepspeed_trainer:228) INFO: 43epoch:train:2801-2900batch: iter_time=1.105e-04, loss_ctc=70.643, loss_att=53.745, acc=0.725, loss=58.791, grad_norm=6.522, loss_scale=1.000, learning_rate=6.158e-05, step_time=0.372 [2024-12-10 22:11:09,731] [INFO] [logging.py:129:log_dist] [Rank 0] step=633000, skipped=0, lr=[np.float64(6.157485854339811e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:11:09,731] [INFO] [timer.py:264:stop] epoch=0/micro_step=33000/global_step=33000, RunningAvgSamplesPerSec=44.1916371632923, CurrSamplesPerSec=43.509456739048716, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:11:09,733 (deepspeed_trainer:228) INFO: 43epoch:train:2901-3000batch: iter_time=1.100e-04, loss_ctc=69.898, loss_att=52.966, acc=0.733, loss=58.044, grad_norm=5.538, loss_scale=1.000, learning_rate=6.158e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:11:46,749 (deepspeed_trainer:228) INFO: 43epoch:train:3001-3100batch: iter_time=1.103e-04, loss_ctc=71.047, loss_att=52.247, acc=0.715, loss=57.883, grad_norm=6.203, loss_scale=1.000, learning_rate=6.157e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:12:23,609 (deepspeed_trainer:228) INFO: 43epoch:train:3101-3200batch: iter_time=1.101e-04, loss_ctc=69.687, loss_att=47.008, acc=0.737, loss=53.806, grad_norm=5.700, loss_scale=1.000, learning_rate=6.157e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:13:00,567 (deepspeed_trainer:228) INFO: 43epoch:train:3201-3300batch: iter_time=1.102e-04, loss_ctc=73.570, loss_att=53.430, acc=0.719, loss=59.501, grad_norm=6.489, loss_scale=1.000, learning_rate=6.156e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:13:37,591 (deepspeed_trainer:228) INFO: 43epoch:train:3301-3400batch: iter_time=1.116e-04, loss_ctc=76.034, loss_att=54.444, acc=0.727, loss=60.925, grad_norm=5.969, loss_scale=1.000, learning_rate=6.156e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:14:14,842 (deepspeed_trainer:228) INFO: 43epoch:train:3401-3500batch: iter_time=1.128e-04, loss_ctc=76.198, loss_att=57.699, acc=0.731, loss=63.244, grad_norm=5.975, loss_scale=1.000, learning_rate=6.155e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:14:51,665 (deepspeed_trainer:228) INFO: 43epoch:train:3501-3600batch: iter_time=1.111e-04, loss_ctc=64.093, loss_att=48.699, acc=0.731, loss=53.336, grad_norm=5.769, loss_scale=1.000, learning_rate=6.155e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:15:28,328 (deepspeed_trainer:228) INFO: 43epoch:train:3601-3700batch: iter_time=1.129e-04, loss_ctc=64.316, loss_att=50.767, acc=0.731, loss=54.823, grad_norm=5.277, loss_scale=1.000, learning_rate=6.154e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:15:50,194 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-018:0/16] 2024-12-10 22:16:16,123 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 22:16:32,635 (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: ) [cnode7-018:0/16] 2024-12-10 22:16:32,636 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 22:16:32,638 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 22:16:58,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:16:59,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:16:59,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:16:59,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,133] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:00,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:01,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:01,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:01,517] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:01,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:47,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,604] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,178] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,855] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:48,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:49,693] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:49,807] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:50,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:49,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:49,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:50,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:50,643] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:17:51,706] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:35,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:36,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:36,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:37,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:37,302] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:37,409] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:38,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:37,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:38,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:38,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:38,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:38,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:39,079] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:39,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:39,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:18:42,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:23,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:24,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:24,672] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:24,992] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:24,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:25,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:25,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:26,032] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:26,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:26,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:26,921] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:26,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:27,193] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:27,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:31,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:19:31,276] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 22:20:16,409 (deepspeed_trainer:228) INFO: 43epoch:train:3701-3800batch: iter_time=2.470, loss_ctc=67.488, loss_att=53.207, acc=0.736, loss=57.494, grad_norm=6.009, loss_scale=1.000, learning_rate=6.154e-05, step_time=0.410 [cnode7-018:0/16] 2024-12-10 22:20:53,541 (deepspeed_trainer:228) INFO: 43epoch:train:3801-3900batch: iter_time=1.077e-04, loss_ctc=71.895, loss_att=56.615, acc=0.726, loss=61.225, grad_norm=6.735, loss_scale=1.000, learning_rate=6.153e-05, step_time=0.371 [2024-12-10 22:21:30,833] [INFO] [logging.py:129:log_dist] [Rank 0] step=634000, skipped=0, lr=[np.float64(6.152627884243594e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:21:30,834] [INFO] [timer.py:264:stop] epoch=0/micro_step=34000/global_step=34000, RunningAvgSamplesPerSec=44.182678038954684, CurrSamplesPerSec=44.62722255721616, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:21:30,836 (deepspeed_trainer:228) INFO: 43epoch:train:3901-4000batch: iter_time=1.073e-04, loss_ctc=75.090, loss_att=57.466, acc=0.721, loss=62.773, grad_norm=6.118, loss_scale=1.000, learning_rate=6.153e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 22:22:07,821 (deepspeed_trainer:228) INFO: 43epoch:train:4001-4100batch: iter_time=1.116e-04, loss_ctc=72.243, loss_att=51.976, acc=0.738, loss=58.073, grad_norm=6.114, loss_scale=1.000, learning_rate=6.152e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:22:44,482 (deepspeed_trainer:228) INFO: 43epoch:train:4101-4200batch: iter_time=1.195e-04, loss_ctc=68.325, loss_att=51.722, acc=0.731, loss=56.746, grad_norm=6.856, loss_scale=1.000, learning_rate=6.152e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 22:23:21,291 (deepspeed_trainer:228) INFO: 43epoch:train:4201-4300batch: iter_time=1.127e-04, loss_ctc=61.765, loss_att=42.465, acc=0.743, loss=48.243, grad_norm=5.853, loss_scale=1.000, learning_rate=6.151e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:23:58,001 (deepspeed_trainer:228) INFO: 43epoch:train:4301-4400batch: iter_time=1.142e-04, loss_ctc=68.792, loss_att=51.994, acc=0.727, loss=57.014, grad_norm=6.656, loss_scale=1.000, learning_rate=6.151e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:24:35,392 (deepspeed_trainer:228) INFO: 43epoch:train:4401-4500batch: iter_time=1.139e-04, loss_ctc=62.495, loss_att=42.695, acc=0.750, loss=48.652, grad_norm=5.603, loss_scale=1.000, learning_rate=6.150e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 22:25:12,538 (deepspeed_trainer:228) INFO: 43epoch:train:4501-4600batch: iter_time=1.139e-04, loss_ctc=71.081, loss_att=51.857, acc=0.739, loss=57.613, grad_norm=6.482, loss_scale=1.000, learning_rate=6.150e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 22:25:49,521 (deepspeed_trainer:228) INFO: 43epoch:train:4601-4700batch: iter_time=1.110e-04, loss_ctc=70.582, loss_att=51.898, acc=0.725, loss=57.505, grad_norm=6.281, loss_scale=1.000, learning_rate=6.149e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:26:26,468 (deepspeed_trainer:228) INFO: 43epoch:train:4701-4800batch: iter_time=1.107e-04, loss_ctc=68.147, loss_att=51.807, acc=0.735, loss=56.725, grad_norm=5.536, loss_scale=1.000, learning_rate=6.149e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:27:03,644 (deepspeed_trainer:228) INFO: 43epoch:train:4801-4900batch: iter_time=1.140e-04, loss_ctc=74.497, loss_att=55.588, acc=0.735, loss=61.256, grad_norm=6.045, loss_scale=1.000, learning_rate=6.149e-05, step_time=0.371 [2024-12-10 22:27:40,428] [INFO] [logging.py:129:log_dist] [Rank 0] step=635000, skipped=0, lr=[np.float64(6.147781394171875e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:27:40,428] [INFO] [timer.py:264:stop] epoch=0/micro_step=35000/global_step=35000, RunningAvgSamplesPerSec=44.189582443140374, CurrSamplesPerSec=46.55600722458164, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:27:40,430 (deepspeed_trainer:228) INFO: 43epoch:train:4901-5000batch: iter_time=1.138e-04, loss_ctc=65.926, loss_att=46.849, acc=0.735, loss=52.578, grad_norm=6.046, loss_scale=1.000, learning_rate=6.148e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:28:17,252 (deepspeed_trainer:228) INFO: 43epoch:train:5001-5100batch: iter_time=1.144e-04, loss_ctc=71.696, loss_att=50.082, acc=0.735, loss=56.566, grad_norm=5.929, loss_scale=1.000, learning_rate=6.148e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 22:28:54,277 (deepspeed_trainer:228) INFO: 43epoch:train:5101-5200batch: iter_time=1.131e-04, loss_ctc=77.316, loss_att=56.201, acc=0.722, loss=62.549, grad_norm=6.357, loss_scale=1.000, learning_rate=6.147e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:29:31,181 (deepspeed_trainer:228) INFO: 43epoch:train:5201-5300batch: iter_time=1.093e-04, loss_ctc=74.099, loss_att=53.816, acc=0.738, loss=59.878, grad_norm=6.190, loss_scale=1.000, learning_rate=6.147e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:30:08,177 (deepspeed_trainer:228) INFO: 43epoch:train:5301-5400batch: iter_time=1.135e-04, loss_ctc=73.081, loss_att=53.612, acc=0.743, loss=59.454, grad_norm=5.779, loss_scale=1.000, learning_rate=6.146e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:30:45,686 (deepspeed_trainer:228) INFO: 43epoch:train:5401-5500batch: iter_time=1.169e-04, loss_ctc=64.319, loss_att=51.157, acc=0.733, loss=55.123, grad_norm=5.388, loss_scale=1.000, learning_rate=6.146e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 22:31:22,595 (deepspeed_trainer:228) INFO: 43epoch:train:5501-5600batch: iter_time=1.136e-04, loss_ctc=63.717, loss_att=49.148, acc=0.734, loss=53.524, grad_norm=5.424, loss_scale=1.000, learning_rate=6.145e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:31:35,636 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-018:0/16] 2024-12-10 22:32:02,306 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 22:32:18,768 (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: ) [cnode7-018:0/16] 2024-12-10 22:32:18,769 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 22:32:18,771 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 22:32:45,671] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:45,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,057] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,552] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,664] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:46,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,723] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:48,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:47,944] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:32:48,073] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:33,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:34,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:35,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:36,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:36,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:36,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,512] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,319] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,526] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:37,852] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:33:39,754] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:21,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:22,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:23,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:24,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:25,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:25,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:26,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:25,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:26,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:26,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:27,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:27,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:27,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:27,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:27,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:34:31,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:08,965] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:11,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:12,230] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:12,312] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:14,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:14,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:14,480] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:14,541] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:14,750] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:15,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:15,343] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:15,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:15,322] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:16,485] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:16,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:35:20,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 22:36:14,313 (deepspeed_trainer:228) INFO: 43epoch:train:5601-5700batch: iter_time=2.491, loss_ctc=67.405, loss_att=49.990, acc=0.738, loss=55.248, grad_norm=5.911, loss_scale=1.000, learning_rate=6.145e-05, step_time=0.426 [cnode7-018:0/16] 2024-12-10 22:36:52,183 (deepspeed_trainer:228) INFO: 43epoch:train:5701-5800batch: iter_time=1.086e-04, loss_ctc=69.075, loss_att=55.662, acc=0.729, loss=59.690, grad_norm=5.771, loss_scale=1.000, learning_rate=6.144e-05, step_time=0.378 [cnode7-018:0/16] 2024-12-10 22:37:30,257 (deepspeed_trainer:228) INFO: 43epoch:train:5801-5900batch: iter_time=1.095e-04, loss_ctc=78.920, loss_att=58.792, acc=0.728, loss=64.805, grad_norm=6.144, loss_scale=1.000, learning_rate=6.144e-05, step_time=0.380 [2024-12-10 22:38:07,746] [INFO] [logging.py:129:log_dist] [Rank 0] step=636000, skipped=0, lr=[np.float64(6.14294633898108e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:38:07,746] [INFO] [timer.py:264:stop] epoch=0/micro_step=36000/global_step=36000, RunningAvgSamplesPerSec=44.16646853812231, CurrSamplesPerSec=41.21029027785074, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:38:07,748 (deepspeed_trainer:228) INFO: 43epoch:train:5901-6000batch: iter_time=1.131e-04, loss_ctc=67.081, loss_att=49.049, acc=0.732, loss=54.481, grad_norm=6.499, loss_scale=1.000, learning_rate=6.143e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 22:38:45,098 (deepspeed_trainer:228) INFO: 43epoch:train:6001-6100batch: iter_time=1.120e-04, loss_ctc=72.504, loss_att=51.691, acc=0.736, loss=57.936, grad_norm=7.460, loss_scale=1.000, learning_rate=6.143e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 22:39:22,333 (deepspeed_trainer:228) INFO: 43epoch:train:6101-6200batch: iter_time=1.173e-04, loss_ctc=61.186, loss_att=43.485, acc=0.735, loss=48.776, grad_norm=5.644, loss_scale=1.000, learning_rate=6.142e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:39:59,576 (deepspeed_trainer:228) INFO: 43epoch:train:6201-6300batch: iter_time=1.154e-04, loss_ctc=67.712, loss_att=51.336, acc=0.734, loss=56.245, grad_norm=6.291, loss_scale=1.000, learning_rate=6.142e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:40:36,184 (deepspeed_trainer:228) INFO: 43epoch:train:6301-6400batch: iter_time=1.153e-04, loss_ctc=63.416, loss_att=46.436, acc=0.745, loss=51.532, grad_norm=5.157, loss_scale=1.000, learning_rate=6.141e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 22:41:13,848 (deepspeed_trainer:228) INFO: 43epoch:train:6401-6500batch: iter_time=1.179e-04, loss_ctc=76.745, loss_att=52.760, acc=0.731, loss=59.967, grad_norm=7.579, loss_scale=1.000, learning_rate=6.141e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-10 22:41:51,166 (deepspeed_trainer:228) INFO: 43epoch:train:6501-6600batch: iter_time=1.147e-04, loss_ctc=70.562, loss_att=53.471, acc=0.724, loss=58.639, grad_norm=6.210, loss_scale=1.000, learning_rate=6.140e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 22:42:28,406 (deepspeed_trainer:228) INFO: 43epoch:train:6601-6700batch: iter_time=1.134e-04, loss_ctc=65.734, loss_att=50.288, acc=0.744, loss=54.920, grad_norm=5.436, loss_scale=1.000, learning_rate=6.140e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:43:05,359 (deepspeed_trainer:228) INFO: 43epoch:train:6701-6800batch: iter_time=1.107e-04, loss_ctc=67.402, loss_att=49.422, acc=0.735, loss=54.807, grad_norm=5.457, loss_scale=1.000, learning_rate=6.139e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:43:42,106 (deepspeed_trainer:228) INFO: 43epoch:train:6801-6900batch: iter_time=1.127e-04, loss_ctc=68.003, loss_att=48.483, acc=0.732, loss=54.332, grad_norm=6.028, loss_scale=1.000, learning_rate=6.139e-05, step_time=0.368 [2024-12-10 22:44:18,879] [INFO] [logging.py:129:log_dist] [Rank 0] step=637000, skipped=0, lr=[np.float64(6.138122673775779e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:44:18,880] [INFO] [timer.py:264:stop] epoch=0/micro_step=37000/global_step=37000, RunningAvgSamplesPerSec=44.16779074899712, CurrSamplesPerSec=46.636049105539314, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:44:18,881 (deepspeed_trainer:228) INFO: 43epoch:train:6901-7000batch: iter_time=1.128e-04, loss_ctc=75.634, loss_att=53.073, acc=0.735, loss=59.825, grad_norm=6.356, loss_scale=1.000, learning_rate=6.138e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:44:56,436 (deepspeed_trainer:228) INFO: 43epoch:train:7001-7100batch: iter_time=1.103e-04, loss_ctc=78.664, loss_att=57.682, acc=0.726, loss=63.967, grad_norm=5.913, loss_scale=1.000, learning_rate=6.138e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 22:45:33,566 (deepspeed_trainer:228) INFO: 43epoch:train:7101-7200batch: iter_time=1.117e-04, loss_ctc=74.757, loss_att=53.441, acc=0.739, loss=59.821, grad_norm=6.224, loss_scale=1.000, learning_rate=6.137e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:46:10,487 (deepspeed_trainer:228) INFO: 43epoch:train:7201-7300batch: iter_time=1.107e-04, loss_ctc=63.538, loss_att=47.572, acc=0.744, loss=52.371, grad_norm=5.565, loss_scale=1.000, learning_rate=6.137e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:46:47,507 (deepspeed_trainer:228) INFO: 43epoch:train:7301-7400batch: iter_time=1.121e-04, loss_ctc=65.945, loss_att=54.480, acc=0.729, loss=57.943, grad_norm=6.168, loss_scale=1.000, learning_rate=6.136e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:47:25,955 (deepspeed_trainer:228) INFO: 43epoch:train:7401-7500batch: iter_time=1.109e-04, loss_ctc=64.652, loss_att=49.089, acc=0.737, loss=53.786, grad_norm=5.775, loss_scale=1.000, learning_rate=6.136e-05, step_time=0.384 [cnode7-018:0/16] 2024-12-10 22:47:30,229 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-018:0/16] 2024-12-10 22:47:57,821 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 22:48:15,061 (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: ) [cnode7-018:0/16] 2024-12-10 22:48:15,062 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 22:48:15,064 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 22:48:40,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:40,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:41,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:41,385] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:41,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:42,005] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:42,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:42,342] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:43,778] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,217] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,467] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,479] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,522] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:48:44,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:29,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:29,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:29,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:30,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:31,353] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:31,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:31,484] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:31,509] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:32,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:32,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:33,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:33,625] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:34,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:34,124] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:34,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:49:34,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:18,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:19,052] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:19,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:19,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:20,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:20,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:21,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:20,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:21,412] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:22,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:22,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:23,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:23,537] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:23,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:23,836] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:50:23,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:07,392] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:08,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:08,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:08,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:08,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:09,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:09,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:09,755] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:09,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:10,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:11,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:11,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:12,413] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:12,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:12,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 22:51:14,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 22:52:18,203 (deepspeed_trainer:228) INFO: 43epoch:train:7501-7600batch: iter_time=2.523, loss_ctc=67.823, loss_att=52.612, acc=0.725, loss=57.153, grad_norm=6.208, loss_scale=1.000, learning_rate=6.135e-05, step_time=0.399 [cnode7-018:0/16] 2024-12-10 22:52:55,638 (deepspeed_trainer:228) INFO: 43epoch:train:7601-7700batch: iter_time=1.125e-04, loss_ctc=72.376, loss_att=59.155, acc=0.716, loss=63.139, grad_norm=7.092, loss_scale=1.000, learning_rate=6.135e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 22:53:33,368 (deepspeed_trainer:228) INFO: 43epoch:train:7701-7800batch: iter_time=1.177e-04, loss_ctc=76.605, loss_att=57.024, acc=0.725, loss=62.911, grad_norm=6.250, loss_scale=1.000, learning_rate=6.135e-05, step_time=0.377 [cnode7-018:0/16] 2024-12-10 22:54:10,239 (deepspeed_trainer:228) INFO: 43epoch:train:7801-7900batch: iter_time=1.159e-04, loss_ctc=68.808, loss_att=49.901, acc=0.733, loss=55.586, grad_norm=6.097, loss_scale=1.000, learning_rate=6.134e-05, step_time=0.368 [2024-12-10 22:54:47,252] [INFO] [logging.py:129:log_dist] [Rank 0] step=638000, skipped=0, lr=[np.float64(6.133310353906922e-05)], mom=[[0.9, 0.98]] [2024-12-10 22:54:47,252] [INFO] [timer.py:264:stop] epoch=0/micro_step=38000/global_step=38000, RunningAvgSamplesPerSec=44.15341157645201, CurrSamplesPerSec=46.480070208707765, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 22:54:47,254 (deepspeed_trainer:228) INFO: 43epoch:train:7901-8000batch: iter_time=1.122e-04, loss_ctc=67.936, loss_att=48.903, acc=0.725, loss=54.632, grad_norm=7.462, loss_scale=1.000, learning_rate=6.134e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:55:24,258 (deepspeed_trainer:228) INFO: 43epoch:train:8001-8100batch: iter_time=1.123e-04, loss_ctc=59.965, loss_att=42.864, acc=0.732, loss=47.998, grad_norm=5.806, loss_scale=1.000, learning_rate=6.133e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 22:56:01,183 (deepspeed_trainer:228) INFO: 43epoch:train:8101-8200batch: iter_time=1.102e-04, loss_ctc=64.200, loss_att=49.964, acc=0.736, loss=54.230, grad_norm=6.272, loss_scale=1.000, learning_rate=6.133e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 22:56:38,391 (deepspeed_trainer:228) INFO: 43epoch:train:8201-8300batch: iter_time=1.122e-04, loss_ctc=75.663, loss_att=54.757, acc=0.733, loss=61.057, grad_norm=6.458, loss_scale=1.000, learning_rate=6.132e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 22:57:15,095 (deepspeed_trainer:228) INFO: 43epoch:train:8301-8400batch: iter_time=1.111e-04, loss_ctc=64.771, loss_att=44.691, acc=0.729, loss=50.724, grad_norm=6.134, loss_scale=1.000, learning_rate=6.132e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:57:51,807 (deepspeed_trainer:228) INFO: 43epoch:train:8401-8500batch: iter_time=1.106e-04, loss_ctc=73.003, loss_att=56.916, acc=0.720, loss=61.724, grad_norm=6.497, loss_scale=1.000, learning_rate=6.131e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 22:58:28,172 (deepspeed_trainer:228) INFO: 43epoch:train:8501-8600batch: iter_time=1.111e-04, loss_ctc=65.918, loss_att=48.637, acc=0.743, loss=53.853, grad_norm=5.744, loss_scale=1.000, learning_rate=6.131e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 22:59:04,559 (deepspeed_trainer:228) INFO: 43epoch:train:8601-8700batch: iter_time=1.130e-04, loss_ctc=68.280, loss_att=50.680, acc=0.722, loss=55.960, grad_norm=6.003, loss_scale=1.000, learning_rate=6.130e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 22:59:41,211 (deepspeed_trainer:228) INFO: 43epoch:train:8701-8800batch: iter_time=1.154e-04, loss_ctc=70.634, loss_att=48.900, acc=0.737, loss=55.452, grad_norm=6.813, loss_scale=1.000, learning_rate=6.130e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:00:18,181 (deepspeed_trainer:228) INFO: 43epoch:train:8801-8900batch: iter_time=1.178e-04, loss_ctc=73.002, loss_att=51.364, acc=0.729, loss=57.813, grad_norm=6.013, loss_scale=1.000, learning_rate=6.129e-05, step_time=0.369 [2024-12-10 23:00:54,272] [INFO] [logging.py:129:log_dist] [Rank 0] step=639000, skipped=0, lr=[np.float64(6.128509334970117e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:00:54,273] [INFO] [timer.py:264:stop] epoch=0/micro_step=39000/global_step=39000, RunningAvgSamplesPerSec=44.16798269997278, CurrSamplesPerSec=47.725076344234296, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:00:54,274 (deepspeed_trainer:228) INFO: 43epoch:train:8901-9000batch: iter_time=1.113e-04, loss_ctc=76.853, loss_att=56.381, acc=0.727, loss=62.515, grad_norm=6.414, loss_scale=1.000, learning_rate=6.129e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 23:01:31,386 (deepspeed_trainer:228) INFO: 43epoch:train:9001-9100batch: iter_time=1.129e-04, loss_ctc=77.696, loss_att=56.438, acc=0.732, loss=62.805, grad_norm=5.960, loss_scale=1.000, learning_rate=6.128e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:02:07,881 (deepspeed_trainer:228) INFO: 43epoch:train:9101-9200batch: iter_time=1.111e-04, loss_ctc=61.000, loss_att=46.607, acc=0.737, loss=50.903, grad_norm=5.865, loss_scale=1.000, learning_rate=6.128e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 23:02:44,453 (deepspeed_trainer:228) INFO: 43epoch:train:9201-9300batch: iter_time=1.108e-04, loss_ctc=63.663, loss_att=50.294, acc=0.730, loss=54.313, grad_norm=6.142, loss_scale=1.000, learning_rate=6.127e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:03:14,732 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-018:0/16] 2024-12-10 23:03:42,282 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 23:03:57,890 (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: ) [cnode7-018:0/16] 2024-12-10 23:03:57,890 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 23:03:57,893 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 23:04:26,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:26,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:26,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:26,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:26,650] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,617] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:27,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:28,484] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:28,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:28,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:04:28,814] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:15,028] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:15,616] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:15,825] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:15,962] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:16,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:16,609] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:16,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,148] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,272] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,330] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,383] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:17,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:18,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:18,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:05:20,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:03,403] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:03,566] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:04,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:05,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:04,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:05,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:05,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:05,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:05,816] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:06,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:06,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:06,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:07,218] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:06,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:07,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:11,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:51,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:51,435] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:53,552] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:53,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:54,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:53,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:54,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:54,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:54,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:55,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:55,184] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:55,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:55,917] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:55,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:06:56,395] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:07:01,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 23:07:35,874 (deepspeed_trainer:228) INFO: 43epoch:train:9301-9400batch: iter_time=2.492, loss_ctc=68.404, loss_att=53.567, acc=0.734, loss=58.012, grad_norm=5.747, loss_scale=1.000, learning_rate=6.127e-05, step_time=0.422 [cnode7-018:0/16] 2024-12-10 23:08:12,661 (deepspeed_trainer:228) INFO: 43epoch:train:9401-9500batch: iter_time=1.102e-04, loss_ctc=70.420, loss_att=54.625, acc=0.721, loss=59.371, grad_norm=6.390, loss_scale=1.000, learning_rate=6.126e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:08:49,508 (deepspeed_trainer:228) INFO: 43epoch:train:9501-9600batch: iter_time=1.106e-04, loss_ctc=72.334, loss_att=55.513, acc=0.720, loss=60.574, grad_norm=6.116, loss_scale=1.000, learning_rate=6.126e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:09:26,278 (deepspeed_trainer:228) INFO: 43epoch:train:9601-9700batch: iter_time=1.138e-04, loss_ctc=74.391, loss_att=52.344, acc=0.739, loss=58.959, grad_norm=6.152, loss_scale=1.000, learning_rate=6.125e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:10:02,836 (deepspeed_trainer:228) INFO: 43epoch:train:9701-9800batch: iter_time=1.093e-04, loss_ctc=67.483, loss_att=49.227, acc=0.725, loss=54.682, grad_norm=6.230, loss_scale=1.000, learning_rate=6.125e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:10:39,307 (deepspeed_trainer:228) INFO: 43epoch:train:9801-9900batch: iter_time=1.092e-04, loss_ctc=67.450, loss_att=48.316, acc=0.728, loss=54.072, grad_norm=6.593, loss_scale=1.000, learning_rate=6.124e-05, step_time=0.364 [2024-12-10 23:11:15,692] [INFO] [logging.py:129:log_dist] [Rank 0] step=640000, skipped=0, lr=[np.float64(6.123719572803898e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:11:15,693] [INFO] [timer.py:264:stop] epoch=0/micro_step=40000/global_step=40000, RunningAvgSamplesPerSec=44.16611074477491, CurrSamplesPerSec=44.343280985601325, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:11:15,695 (deepspeed_trainer:228) INFO: 43epoch:train:9901-10000batch: iter_time=1.105e-04, loss_ctc=64.584, loss_att=47.889, acc=0.728, loss=52.907, grad_norm=7.013, loss_scale=1.000, learning_rate=6.124e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-10 23:11:52,346 (deepspeed_trainer:228) INFO: 43epoch:train:10001-10100batch: iter_time=1.103e-04, loss_ctc=62.873, loss_att=44.553, acc=0.739, loss=50.042, grad_norm=6.183, loss_scale=1.000, learning_rate=6.123e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:12:29,141 (deepspeed_trainer:228) INFO: 43epoch:train:10101-10200batch: iter_time=1.109e-04, loss_ctc=73.246, loss_att=54.112, acc=0.737, loss=59.823, grad_norm=6.559, loss_scale=1.000, learning_rate=6.123e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:13:05,786 (deepspeed_trainer:228) INFO: 43epoch:train:10201-10300batch: iter_time=1.148e-04, loss_ctc=67.001, loss_att=47.050, acc=0.721, loss=53.069, grad_norm=6.147, loss_scale=1.000, learning_rate=6.123e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:13:42,740 (deepspeed_trainer:228) INFO: 43epoch:train:10301-10400batch: iter_time=1.160e-04, loss_ctc=69.726, loss_att=53.049, acc=0.731, loss=58.025, grad_norm=6.087, loss_scale=1.000, learning_rate=6.122e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:14:19,851 (deepspeed_trainer:228) INFO: 43epoch:train:10401-10500batch: iter_time=1.148e-04, loss_ctc=69.259, loss_att=52.501, acc=0.738, loss=57.518, grad_norm=6.165, loss_scale=1.000, learning_rate=6.122e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:14:56,794 (deepspeed_trainer:228) INFO: 43epoch:train:10501-10600batch: iter_time=1.130e-04, loss_ctc=70.386, loss_att=51.453, acc=0.721, loss=57.110, grad_norm=6.272, loss_scale=1.000, learning_rate=6.121e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:15:33,677 (deepspeed_trainer:228) INFO: 43epoch:train:10601-10700batch: iter_time=1.149e-04, loss_ctc=68.754, loss_att=46.561, acc=0.739, loss=53.230, grad_norm=5.650, loss_scale=1.000, learning_rate=6.121e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:16:10,423 (deepspeed_trainer:228) INFO: 43epoch:train:10701-10800batch: iter_time=1.158e-04, loss_ctc=72.993, loss_att=52.293, acc=0.725, loss=58.524, grad_norm=6.629, loss_scale=1.000, learning_rate=6.120e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:16:47,338 (deepspeed_trainer:228) INFO: 43epoch:train:10801-10900batch: iter_time=1.096e-04, loss_ctc=74.649, loss_att=53.478, acc=0.733, loss=59.826, grad_norm=6.478, loss_scale=1.000, learning_rate=6.120e-05, step_time=0.369 [2024-12-10 23:17:24,452] [INFO] [logging.py:129:log_dist] [Rank 0] step=641000, skipped=0, lr=[np.float64(6.118941023488029e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:17:24,453] [INFO] [timer.py:264:stop] epoch=0/micro_step=41000/global_step=41000, RunningAvgSamplesPerSec=44.17456645097412, CurrSamplesPerSec=46.25326005431564, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:17:24,455 (deepspeed_trainer:228) INFO: 43epoch:train:10901-11000batch: iter_time=1.148e-04, loss_ctc=74.914, loss_att=56.487, acc=0.734, loss=62.002, grad_norm=5.725, loss_scale=1.000, learning_rate=6.119e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:18:01,198 (deepspeed_trainer:228) INFO: 43epoch:train:11001-11100batch: iter_time=1.153e-04, loss_ctc=63.659, loss_att=48.598, acc=0.732, loss=53.101, grad_norm=5.524, loss_scale=1.000, learning_rate=6.119e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:18:38,249 (deepspeed_trainer:228) INFO: 43epoch:train:11101-11200batch: iter_time=1.118e-04, loss_ctc=65.136, loss_att=50.597, acc=0.731, loss=54.933, grad_norm=6.114, loss_scale=1.000, learning_rate=6.118e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:19:00,195 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-018:0/16] 2024-12-10 23:19:27,667 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 23:19:43,896 (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: ) [cnode7-018:0/16] 2024-12-10 23:19:43,896 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 23:19:43,899 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 23:20:13,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:12,903] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,648] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,235] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,870] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,583] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,587] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:14,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:14,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:14,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:14,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:20:13,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:02,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:01,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:01,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:02,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:02,582] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,418] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:02,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,290] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:04,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:04,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:03,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:50,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:50,602] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:50,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:51,361] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:51,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:51,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:53,121] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,677] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:53,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:53,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:52,972] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:21:54,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:38,504] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:39,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:39,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:39,796] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:40,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:40,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:40,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:40,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:41,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:44,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:22:45,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 23:23:30,498 (deepspeed_trainer:228) INFO: 43epoch:train:11201-11300batch: iter_time=2.507, loss_ctc=67.123, loss_att=51.947, acc=0.736, loss=56.481, grad_norm=5.445, loss_scale=1.000, learning_rate=6.118e-05, step_time=0.414 [cnode7-018:0/16] 2024-12-10 23:24:07,287 (deepspeed_trainer:228) INFO: 43epoch:train:11301-11400batch: iter_time=1.055e-04, loss_ctc=71.537, loss_att=55.155, acc=0.724, loss=60.082, grad_norm=6.259, loss_scale=1.000, learning_rate=6.117e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:24:43,787 (deepspeed_trainer:228) INFO: 43epoch:train:11401-11500batch: iter_time=1.077e-04, loss_ctc=73.861, loss_att=55.704, acc=0.717, loss=61.187, grad_norm=6.341, loss_scale=1.000, learning_rate=6.117e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-10 23:25:20,234 (deepspeed_trainer:228) INFO: 43epoch:train:11501-11600batch: iter_time=1.102e-04, loss_ctc=71.969, loss_att=51.303, acc=0.736, loss=57.506, grad_norm=6.292, loss_scale=1.000, learning_rate=6.116e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 23:25:56,449 (deepspeed_trainer:228) INFO: 43epoch:train:11601-11700batch: iter_time=1.095e-04, loss_ctc=67.409, loss_att=50.677, acc=0.728, loss=55.662, grad_norm=7.410, loss_scale=1.000, learning_rate=6.116e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-10 23:26:32,570 (deepspeed_trainer:228) INFO: 43epoch:train:11701-11800batch: iter_time=1.077e-04, loss_ctc=61.031, loss_att=41.706, acc=0.741, loss=47.495, grad_norm=5.970, loss_scale=1.000, learning_rate=6.115e-05, step_time=0.361 [cnode7-018:0/16] 2024-12-10 23:27:08,880 (deepspeed_trainer:228) INFO: 43epoch:train:11801-11900batch: iter_time=1.076e-04, loss_ctc=68.308, loss_att=52.929, acc=0.717, loss=57.565, grad_norm=6.664, loss_scale=1.000, learning_rate=6.115e-05, step_time=0.363 [2024-12-10 23:27:45,606] [INFO] [logging.py:129:log_dist] [Rank 0] step=642000, skipped=0, lr=[np.float64(6.114173643341806e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:27:45,607] [INFO] [timer.py:264:stop] epoch=0/micro_step=42000/global_step=42000, RunningAvgSamplesPerSec=44.17821222958442, CurrSamplesPerSec=46.69192400836058, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:27:45,608 (deepspeed_trainer:228) INFO: 43epoch:train:11901-12000batch: iter_time=1.154e-04, loss_ctc=61.300, loss_att=42.828, acc=0.749, loss=48.373, grad_norm=5.302, loss_scale=1.000, learning_rate=6.114e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:28:22,724 (deepspeed_trainer:228) INFO: 43epoch:train:12001-12100batch: iter_time=1.105e-04, loss_ctc=70.435, loss_att=51.182, acc=0.739, loss=56.954, grad_norm=6.656, loss_scale=1.000, learning_rate=6.114e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:28:59,440 (deepspeed_trainer:228) INFO: 43epoch:train:12101-12200batch: iter_time=1.090e-04, loss_ctc=69.619, loss_att=52.195, acc=0.717, loss=57.445, grad_norm=6.448, loss_scale=1.000, learning_rate=6.113e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:29:36,444 (deepspeed_trainer:228) INFO: 43epoch:train:12201-12300batch: iter_time=1.081e-04, loss_ctc=67.744, loss_att=51.028, acc=0.729, loss=56.054, grad_norm=5.908, loss_scale=1.000, learning_rate=6.113e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 23:30:13,960 (deepspeed_trainer:228) INFO: 43epoch:train:12301-12400batch: iter_time=1.098e-04, loss_ctc=73.889, loss_att=53.885, acc=0.730, loss=59.894, grad_norm=6.733, loss_scale=1.000, learning_rate=6.113e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 23:30:51,258 (deepspeed_trainer:228) INFO: 43epoch:train:12401-12500batch: iter_time=1.103e-04, loss_ctc=65.287, loss_att=46.513, acc=0.731, loss=52.158, grad_norm=6.132, loss_scale=1.000, learning_rate=6.112e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 23:31:28,519 (deepspeed_trainer:228) INFO: 43epoch:train:12501-12600batch: iter_time=1.103e-04, loss_ctc=70.698, loss_att=49.271, acc=0.736, loss=55.707, grad_norm=6.270, loss_scale=1.000, learning_rate=6.112e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 23:32:05,966 (deepspeed_trainer:228) INFO: 43epoch:train:12601-12700batch: iter_time=1.092e-04, loss_ctc=76.951, loss_att=56.511, acc=0.718, loss=62.668, grad_norm=6.943, loss_scale=1.000, learning_rate=6.111e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-10 23:32:43,238 (deepspeed_trainer:228) INFO: 43epoch:train:12701-12800batch: iter_time=1.103e-04, loss_ctc=72.866, loss_att=52.787, acc=0.733, loss=58.832, grad_norm=6.290, loss_scale=1.000, learning_rate=6.111e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 23:33:20,736 (deepspeed_trainer:228) INFO: 43epoch:train:12801-12900batch: iter_time=1.115e-04, loss_ctc=72.610, loss_att=53.161, acc=0.740, loss=58.968, grad_norm=6.378, loss_scale=1.000, learning_rate=6.110e-05, step_time=0.375 [2024-12-10 23:33:58,304] [INFO] [logging.py:129:log_dist] [Rank 0] step=643000, skipped=0, lr=[np.float64(6.109417388922379e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:33:58,305] [INFO] [timer.py:264:stop] epoch=0/micro_step=43000/global_step=43000, RunningAvgSamplesPerSec=44.17512142439763, CurrSamplesPerSec=45.241264032845436, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:33:58,306 (deepspeed_trainer:228) INFO: 43epoch:train:12901-13000batch: iter_time=1.103e-04, loss_ctc=63.821, loss_att=50.488, acc=0.732, loss=54.508, grad_norm=5.633, loss_scale=1.000, learning_rate=6.110e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 23:34:35,459 (deepspeed_trainer:228) INFO: 43epoch:train:13001-13100batch: iter_time=1.101e-04, loss_ctc=63.364, loss_att=48.856, acc=0.731, loss=53.192, grad_norm=5.827, loss_scale=1.000, learning_rate=6.109e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:34:48,054 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-018:0/16] 2024-12-10 23:35:15,181 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 23:35:30,903 (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: ) [cnode7-018:0/16] 2024-12-10 23:35:30,903 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 23:35:30,905 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 23:35:58,943] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:00,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:35:59,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:00,205] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:00,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:00,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:00,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:01,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:01,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:01,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:01,647] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:47,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:47,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:48,194] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:48,762] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:49,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:49,465] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:49,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:49,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:49,827] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:50,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:50,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:50,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:50,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:51,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:51,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:36:54,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,063] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:37,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:38,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:38,673] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:38,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:39,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:39,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:39,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:39,325] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:40,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:43,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:37:45,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:25,211] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:25,457] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:25,729] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:26,166] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:27,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:27,714] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:27,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:27,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:27,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:28,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:28,567] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:28,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:28,612] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:29,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:33,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:38:37,456] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 23:39:31,375 (deepspeed_trainer:228) INFO: 43epoch:train:13101-13200batch: iter_time=2.504, loss_ctc=66.812, loss_att=51.296, acc=0.737, loss=55.965, grad_norm=5.910, loss_scale=1.000, learning_rate=6.109e-05, step_time=0.455 [cnode7-018:0/16] 2024-12-10 23:40:08,710 (deepspeed_trainer:228) INFO: 43epoch:train:13201-13300batch: iter_time=1.121e-04, loss_ctc=68.082, loss_att=56.616, acc=0.728, loss=60.058, grad_norm=6.247, loss_scale=1.000, learning_rate=6.108e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-10 23:40:46,274 (deepspeed_trainer:228) INFO: 43epoch:train:13301-13400batch: iter_time=1.105e-04, loss_ctc=78.659, loss_att=59.050, acc=0.728, loss=64.931, grad_norm=6.624, loss_scale=1.000, learning_rate=6.108e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-10 23:41:23,126 (deepspeed_trainer:228) INFO: 43epoch:train:13401-13500batch: iter_time=1.114e-04, loss_ctc=67.131, loss_att=49.656, acc=0.733, loss=54.886, grad_norm=7.286, loss_scale=1.000, learning_rate=6.107e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:42:00,137 (deepspeed_trainer:228) INFO: 43epoch:train:13501-13600batch: iter_time=1.124e-04, loss_ctc=70.773, loss_att=52.310, acc=0.736, loss=57.859, grad_norm=7.861, loss_scale=1.000, learning_rate=6.107e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 23:42:36,795 (deepspeed_trainer:228) INFO: 43epoch:train:13601-13700batch: iter_time=1.118e-04, loss_ctc=60.655, loss_att=43.436, acc=0.738, loss=48.580, grad_norm=5.632, loss_scale=1.000, learning_rate=6.106e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:43:13,817 (deepspeed_trainer:228) INFO: 43epoch:train:13701-13800batch: iter_time=1.105e-04, loss_ctc=66.617, loss_att=50.759, acc=0.737, loss=55.520, grad_norm=6.198, loss_scale=1.000, learning_rate=6.106e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 23:43:50,970 (deepspeed_trainer:228) INFO: 43epoch:train:13801-13900batch: iter_time=1.123e-04, loss_ctc=62.799, loss_att=46.228, acc=0.748, loss=51.212, grad_norm=5.448, loss_scale=1.000, learning_rate=6.105e-05, step_time=0.371 [2024-12-10 23:44:28,183] [INFO] [logging.py:129:log_dist] [Rank 0] step=644000, skipped=0, lr=[np.float64(6.104672217023095e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:44:28,184] [INFO] [timer.py:264:stop] epoch=0/micro_step=44000/global_step=44000, RunningAvgSamplesPerSec=44.15356809651246, CurrSamplesPerSec=44.15768029714401, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:44:28,185 (deepspeed_trainer:228) INFO: 43epoch:train:13901-14000batch: iter_time=1.093e-04, loss_ctc=75.606, loss_att=52.554, acc=0.735, loss=59.447, grad_norm=7.219, loss_scale=1.000, learning_rate=6.105e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 23:45:05,285 (deepspeed_trainer:228) INFO: 43epoch:train:14001-14100batch: iter_time=1.132e-04, loss_ctc=69.326, loss_att=53.728, acc=0.725, loss=58.402, grad_norm=6.617, loss_scale=1.000, learning_rate=6.104e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-10 23:45:42,509 (deepspeed_trainer:228) INFO: 43epoch:train:14101-14200batch: iter_time=1.118e-04, loss_ctc=65.165, loss_att=50.339, acc=0.746, loss=54.798, grad_norm=5.660, loss_scale=1.000, learning_rate=6.104e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-10 23:46:19,472 (deepspeed_trainer:228) INFO: 43epoch:train:14201-14300batch: iter_time=1.111e-04, loss_ctc=67.168, loss_att=49.589, acc=0.734, loss=54.832, grad_norm=6.062, loss_scale=1.000, learning_rate=6.103e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:46:56,054 (deepspeed_trainer:228) INFO: 43epoch:train:14301-14400batch: iter_time=1.098e-04, loss_ctc=67.079, loss_att=48.353, acc=0.736, loss=53.979, grad_norm=5.835, loss_scale=1.000, learning_rate=6.103e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-10 23:47:32,817 (deepspeed_trainer:228) INFO: 43epoch:train:14401-14500batch: iter_time=1.094e-04, loss_ctc=74.794, loss_att=52.511, acc=0.737, loss=59.183, grad_norm=6.283, loss_scale=1.000, learning_rate=6.103e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:48:09,927 (deepspeed_trainer:228) INFO: 43epoch:train:14501-14600batch: iter_time=1.105e-04, loss_ctc=78.904, loss_att=57.682, acc=0.726, loss=64.035, grad_norm=6.875, loss_scale=1.000, learning_rate=6.102e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-10 23:48:46,835 (deepspeed_trainer:228) INFO: 43epoch:train:14601-14700batch: iter_time=1.093e-04, loss_ctc=74.819, loss_att=53.783, acc=0.738, loss=60.090, grad_norm=6.011, loss_scale=1.000, learning_rate=6.102e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:49:23,225 (deepspeed_trainer:228) INFO: 43epoch:train:14701-14800batch: iter_time=1.130e-04, loss_ctc=62.996, loss_att=47.563, acc=0.747, loss=52.192, grad_norm=5.373, loss_scale=1.000, learning_rate=6.101e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-10 23:49:59,795 (deepspeed_trainer:228) INFO: 43epoch:train:14801-14900batch: iter_time=1.116e-04, loss_ctc=66.014, loss_att=54.288, acc=0.730, loss=57.789, grad_norm=5.954, loss_scale=1.000, learning_rate=6.101e-05, step_time=0.365 [2024-12-10 23:50:36,426] [INFO] [logging.py:129:log_dist] [Rank 0] step=645000, skipped=0, lr=[np.float64(6.099938084671846e-05)], mom=[[0.9, 0.98]] [2024-12-10 23:50:36,427] [INFO] [timer.py:264:stop] epoch=0/micro_step=45000/global_step=45000, RunningAvgSamplesPerSec=44.16311484860535, CurrSamplesPerSec=47.76243970489904, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-10 23:50:36,428 (deepspeed_trainer:228) INFO: 43epoch:train:14901-15000batch: iter_time=1.086e-04, loss_ctc=64.478, loss_att=49.094, acc=0.736, loss=53.723, grad_norm=6.144, loss_scale=1.000, learning_rate=6.100e-05, step_time=0.366 [2024-12-10 23:50:49,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:50,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,282] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:51,350] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,330] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,553] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,597] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:50:52,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:05,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:07,024] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:06,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:06,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:07,131] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:07,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:07,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:08,681] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:08,728] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:08,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:08,312] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:08,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:09,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:09,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:09,688] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:09,694] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:20,646] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:22,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:22,449] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:22,978] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:23,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:24,111] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:23,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:24,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:24,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:24,473] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:24,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:25,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:25,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:26,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:26,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:26,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:36,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:37,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:37,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:38,643] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:38,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:39,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:39,417] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:40,624] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:40,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:40,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:41,245] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:41,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:41,129] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:41,951] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:42,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:41,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:51:54,832] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 43 is about to be saved! [2024-12-10 23:51:54,863] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/mp_rank_00_model_states.pt [2024-12-10 23:51:54,863] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/mp_rank_00_model_states.pt... [2024-12-10 23:51:56,772] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/mp_rank_00_model_states.pt. [2024-12-10 23:51:56,928] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,930] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,930] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,162] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,163] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,164] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,932] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,165] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,165] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,165] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,934] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,934] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,168] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,936] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,936] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-10 23:51:56,168] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-10 23:51:57,520] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,520] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,520] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,685] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,692] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,692] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,697] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,697] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,697] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,722] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,722] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,722] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,725] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,725] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,725] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,748] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,748] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,748] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,767] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,767] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,767] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,767] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,767] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,767] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_43/43/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [2024-12-10 23:51:57,309] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 43 is ready now! [cnode7-018:0/16] 2024-12-10 23:51:58,128 (deepspeed_trainer:158) INFO: 43epoch results: [train] iter_time=0.133, loss_ctc=69.532, loss_att=51.483, acc=0.732, loss=56.900, grad_norm=6.184, loss_scale=1.000, learning_rate=6.136e-05, step_time=0.372, time=2 hours, 6 minutes and 25.35 seconds, total_count=645043, gpu_max_cached_mem_GB=28.656, [valid] loss_ctc=3.375, cer_ctc=0.053, loss_att=7.281, acc=0.815, cer=0.418, wer=1.000, loss=6.125, time=1 minute and 10.12 seconds, total_count=43, gpu_max_cached_mem_GB=28.656 [cnode7-018:0/16] 2024-12-10 23:52:01,003 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-018:0/16] 2024-12-10 23:52:27,153 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-10 23:52:43,734 (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: ) [cnode7-018:0/16] 2024-12-10 23:52:43,734 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-10 23:52:43,737 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-10 23:53:02,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:02,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:03,584] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:03,797] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:03,218] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:03,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:04,232] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:03,567] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:04,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:04,622] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:05,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:04,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:05,125] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:05,183] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:05,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:07,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:51,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:50,698] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:50,881] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:52,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:51,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:52,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:51,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:52,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:52,693] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:53,349] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:52,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:53,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:54,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:54,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:54,419] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:53:55,726] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:39,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:38,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:38,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:40,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:39,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:39,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:40,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:39,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:40,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,012] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:41,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:54:46,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:26,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:26,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:26,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:28,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:27,795] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:28,228] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:29,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:28,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:29,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:30,248] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:30,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:30,373] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:29,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:29,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:30,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-10 23:55:35,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-10 23:56:39,715 (deepspeed_trainer:228) INFO: 44epoch:train:1-100batch: iter_time=2.369, loss_ctc=62.964, loss_att=46.410, acc=0.737, loss=51.392, grad_norm=6.036, loss_scale=1.000, learning_rate=6.100e-05, step_time=0.419 [cnode7-018:0/16] 2024-12-10 23:57:16,558 (deepspeed_trainer:228) INFO: 44epoch:train:101-200batch: iter_time=1.097e-04, loss_ctc=68.599, loss_att=53.553, acc=0.724, loss=58.084, grad_norm=6.380, loss_scale=1.000, learning_rate=6.099e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-10 23:57:53,516 (deepspeed_trainer:228) INFO: 44epoch:train:201-300batch: iter_time=1.127e-04, loss_ctc=78.079, loss_att=54.879, acc=0.741, loss=61.828, grad_norm=7.190, loss_scale=1.000, learning_rate=6.099e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:58:30,304 (deepspeed_trainer:228) INFO: 44epoch:train:301-400batch: iter_time=1.106e-04, loss_ctc=70.216, loss_att=55.653, acc=0.713, loss=60.048, grad_norm=6.829, loss_scale=1.000, learning_rate=6.098e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-10 23:59:07,254 (deepspeed_trainer:228) INFO: 44epoch:train:401-500batch: iter_time=1.164e-04, loss_ctc=74.189, loss_att=49.315, acc=0.739, loss=56.785, grad_norm=6.607, loss_scale=1.000, learning_rate=6.098e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-10 23:59:44,017 (deepspeed_trainer:228) INFO: 44epoch:train:501-600batch: iter_time=1.175e-04, loss_ctc=69.145, loss_att=51.772, acc=0.727, loss=57.015, grad_norm=7.707, loss_scale=1.000, learning_rate=6.097e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:00:20,487 (deepspeed_trainer:228) INFO: 44epoch:train:601-700batch: iter_time=1.134e-04, loss_ctc=74.993, loss_att=55.554, acc=0.720, loss=61.364, grad_norm=7.709, loss_scale=1.000, learning_rate=6.097e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:00:57,456 (deepspeed_trainer:228) INFO: 44epoch:train:701-800batch: iter_time=1.159e-04, loss_ctc=61.847, loss_att=44.204, acc=0.736, loss=49.495, grad_norm=5.826, loss_scale=1.000, learning_rate=6.096e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:01:34,569 (deepspeed_trainer:228) INFO: 44epoch:train:801-900batch: iter_time=1.127e-04, loss_ctc=67.712, loss_att=47.328, acc=0.727, loss=53.438, grad_norm=6.993, loss_scale=1.000, learning_rate=6.096e-05, step_time=0.371 [2024-12-11 00:02:11,889] [INFO] [logging.py:129:log_dist] [Rank 0] step=646000, skipped=0, lr=[np.float64(6.095214949129432e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:02:11,890] [INFO] [timer.py:264:stop] epoch=0/micro_step=46000/global_step=46000, RunningAvgSamplesPerSec=44.156439598308154, CurrSamplesPerSec=43.00446562247338, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:02:11,891 (deepspeed_trainer:228) INFO: 44epoch:train:901-1000batch: iter_time=1.177e-04, loss_ctc=74.871, loss_att=56.707, acc=0.712, loss=62.163, grad_norm=6.495, loss_scale=1.000, learning_rate=6.095e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 00:02:48,987 (deepspeed_trainer:228) INFO: 44epoch:train:1001-1100batch: iter_time=1.142e-04, loss_ctc=68.531, loss_att=47.904, acc=0.732, loss=54.102, grad_norm=6.143, loss_scale=1.000, learning_rate=6.095e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:03:26,449 (deepspeed_trainer:228) INFO: 44epoch:train:1101-1200batch: iter_time=1.109e-04, loss_ctc=63.534, loss_att=48.027, acc=0.739, loss=52.681, grad_norm=6.201, loss_scale=1.000, learning_rate=6.095e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 00:04:03,823 (deepspeed_trainer:228) INFO: 44epoch:train:1201-1300batch: iter_time=1.107e-04, loss_ctc=65.713, loss_att=47.563, acc=0.736, loss=52.970, grad_norm=6.213, loss_scale=1.000, learning_rate=6.094e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 00:04:40,778 (deepspeed_trainer:228) INFO: 44epoch:train:1301-1400batch: iter_time=1.120e-04, loss_ctc=74.291, loss_att=53.720, acc=0.715, loss=59.892, grad_norm=7.303, loss_scale=1.000, learning_rate=6.094e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:05:17,759 (deepspeed_trainer:228) INFO: 44epoch:train:1401-1500batch: iter_time=1.116e-04, loss_ctc=74.840, loss_att=49.651, acc=0.725, loss=57.216, grad_norm=7.339, loss_scale=1.000, learning_rate=6.093e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:05:55,303 (deepspeed_trainer:228) INFO: 44epoch:train:1501-1600batch: iter_time=1.103e-04, loss_ctc=70.757, loss_att=56.313, acc=0.726, loss=60.655, grad_norm=6.911, loss_scale=1.000, learning_rate=6.093e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-11 00:06:32,547 (deepspeed_trainer:228) INFO: 44epoch:train:1601-1700batch: iter_time=1.126e-04, loss_ctc=66.561, loss_att=47.023, acc=0.730, loss=52.893, grad_norm=6.658, loss_scale=1.000, learning_rate=6.092e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 00:07:09,901 (deepspeed_trainer:228) INFO: 44epoch:train:1701-1800batch: iter_time=1.117e-04, loss_ctc=75.752, loss_att=59.297, acc=0.733, loss=64.227, grad_norm=6.199, loss_scale=1.000, learning_rate=6.092e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 00:07:41,476 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-018:0/16] 2024-12-11 00:08:07,956 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 00:08:23,730 (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: ) [cnode7-018:0/16] 2024-12-11 00:08:23,731 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 00:08:23,733 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 00:08:55,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:55,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:55,998] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,749] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,811] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:56,995] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:08:57,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:48,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:48,922] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:49,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:49,660] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:49,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:49,844] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,198] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,585] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:50,366] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:51,121] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:51,199] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:51,274] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:09:51,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:36,834] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:37,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:37,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:37,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:38,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:38,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:38,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,022] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,397] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:39,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:40,072] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:43,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:10:43,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:25,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:25,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:26,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:26,256] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:26,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:27,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:27,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:27,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:27,572] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:28,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:28,221] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:28,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:28,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:28,785] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:33,512] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:11:33,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 00:12:09,812 (deepspeed_trainer:228) INFO: 44epoch:train:1801-1900batch: iter_time=2.558, loss_ctc=60.431, loss_att=50.363, acc=0.713, loss=53.384, grad_norm=6.445, loss_scale=1.000, learning_rate=6.091e-05, step_time=0.439 [2024-12-11 00:12:46,891] [INFO] [logging.py:129:log_dist] [Rank 0] step=647000, skipped=0, lr=[np.float64(6.090502767887948e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:12:46,895] [INFO] [timer.py:264:stop] epoch=0/micro_step=47000/global_step=47000, RunningAvgSamplesPerSec=44.13716690611898, CurrSamplesPerSec=40.82027342230548, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:12:46,897 (deepspeed_trainer:228) INFO: 44epoch:train:1901-2000batch: iter_time=1.130e-04, loss_ctc=64.153, loss_att=47.809, acc=0.734, loss=52.703, grad_norm=6.368, loss_scale=1.000, learning_rate=6.091e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:13:24,029 (deepspeed_trainer:228) INFO: 44epoch:train:2001-2100batch: iter_time=1.130e-04, loss_ctc=64.510, loss_att=50.827, acc=0.726, loss=54.935, grad_norm=5.396, loss_scale=1.000, learning_rate=6.090e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:14:01,297 (deepspeed_trainer:228) INFO: 44epoch:train:2101-2200batch: iter_time=1.124e-04, loss_ctc=75.692, loss_att=57.986, acc=0.727, loss=63.299, grad_norm=7.649, loss_scale=1.000, learning_rate=6.090e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 00:14:38,147 (deepspeed_trainer:228) INFO: 44epoch:train:2201-2300batch: iter_time=1.140e-04, loss_ctc=69.584, loss_att=49.502, acc=0.721, loss=55.528, grad_norm=6.610, loss_scale=1.000, learning_rate=6.089e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 00:15:15,205 (deepspeed_trainer:228) INFO: 44epoch:train:2301-2400batch: iter_time=1.116e-04, loss_ctc=71.529, loss_att=49.081, acc=0.732, loss=55.835, grad_norm=6.267, loss_scale=1.000, learning_rate=6.089e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:15:52,308 (deepspeed_trainer:228) INFO: 44epoch:train:2401-2500batch: iter_time=1.112e-04, loss_ctc=69.204, loss_att=50.257, acc=0.726, loss=55.930, grad_norm=6.767, loss_scale=1.000, learning_rate=6.088e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:16:29,375 (deepspeed_trainer:228) INFO: 44epoch:train:2501-2600batch: iter_time=1.166e-04, loss_ctc=69.379, loss_att=52.413, acc=0.728, loss=57.530, grad_norm=7.244, loss_scale=1.000, learning_rate=6.088e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:17:05,908 (deepspeed_trainer:228) INFO: 44epoch:train:2601-2700batch: iter_time=1.156e-04, loss_ctc=62.483, loss_att=42.216, acc=0.737, loss=48.307, grad_norm=5.847, loss_scale=1.000, learning_rate=6.087e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 00:17:42,385 (deepspeed_trainer:228) INFO: 44epoch:train:2701-2800batch: iter_time=1.151e-04, loss_ctc=69.820, loss_att=50.001, acc=0.722, loss=55.966, grad_norm=6.523, loss_scale=1.000, learning_rate=6.087e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 00:18:19,233 (deepspeed_trainer:228) INFO: 44epoch:train:2801-2900batch: iter_time=1.096e-04, loss_ctc=74.458, loss_att=55.713, acc=0.711, loss=61.354, grad_norm=6.788, loss_scale=1.000, learning_rate=6.087e-05, step_time=0.368 [2024-12-11 00:18:55,590] [INFO] [logging.py:129:log_dist] [Rank 0] step=648000, skipped=0, lr=[np.float64(6.0858014986691676e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:18:55,591] [INFO] [timer.py:264:stop] epoch=0/micro_step=48000/global_step=48000, RunningAvgSamplesPerSec=44.1454666685599, CurrSamplesPerSec=46.9581336190895, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:18:55,593 (deepspeed_trainer:228) INFO: 44epoch:train:2901-3000batch: iter_time=1.100e-04, loss_ctc=66.041, loss_att=43.343, acc=0.740, loss=50.142, grad_norm=5.608, loss_scale=1.000, learning_rate=6.086e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 00:19:32,258 (deepspeed_trainer:228) INFO: 44epoch:train:3001-3100batch: iter_time=1.143e-04, loss_ctc=63.368, loss_att=48.822, acc=0.729, loss=53.179, grad_norm=5.928, loss_scale=1.000, learning_rate=6.086e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:20:09,341 (deepspeed_trainer:228) INFO: 44epoch:train:3101-3200batch: iter_time=1.119e-04, loss_ctc=72.601, loss_att=53.226, acc=0.732, loss=59.023, grad_norm=6.705, loss_scale=1.000, learning_rate=6.085e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:20:45,592 (deepspeed_trainer:228) INFO: 44epoch:train:3201-3300batch: iter_time=1.125e-04, loss_ctc=65.573, loss_att=45.722, acc=0.727, loss=51.690, grad_norm=6.442, loss_scale=1.000, learning_rate=6.085e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 00:21:21,965 (deepspeed_trainer:228) INFO: 44epoch:train:3301-3400batch: iter_time=1.080e-04, loss_ctc=69.912, loss_att=48.933, acc=0.719, loss=55.234, grad_norm=7.559, loss_scale=1.000, learning_rate=6.084e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 00:21:58,571 (deepspeed_trainer:228) INFO: 44epoch:train:3401-3500batch: iter_time=1.112e-04, loss_ctc=69.060, loss_att=54.042, acc=0.729, loss=58.516, grad_norm=6.202, loss_scale=1.000, learning_rate=6.084e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:22:35,191 (deepspeed_trainer:228) INFO: 44epoch:train:3501-3600batch: iter_time=1.115e-04, loss_ctc=74.915, loss_att=53.825, acc=0.730, loss=60.160, grad_norm=6.691, loss_scale=1.000, learning_rate=6.083e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 00:23:11,640 (deepspeed_trainer:228) INFO: 44epoch:train:3601-3700batch: iter_time=1.108e-04, loss_ctc=69.134, loss_att=56.596, acc=0.719, loss=60.348, grad_norm=6.615, loss_scale=1.000, learning_rate=6.083e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 00:23:32,962 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-018:0/16] 2024-12-11 00:24:00,233 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 00:24:15,828 (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: ) [cnode7-018:0/16] 2024-12-11 00:24:15,829 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 00:24:15,831 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 00:24:42,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:43,209] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:43,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:43,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:43,957] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:44,002] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:44,563] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:44,676] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:44,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,123] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:44,860] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,207] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:24:45,287] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:32,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:32,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,625] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,548] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:34,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:33,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:34,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:34,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:34,842] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:34,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:25:35,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:20,007] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:20,819] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:21,151] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:21,369] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:21,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:22,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:22,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:22,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:23,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:22,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:23,008] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:23,305] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:23,746] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:24,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:25,782] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:26:26,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:07,426] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:08,843] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:09,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:09,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:09,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:10,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:10,598] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:11,144] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:10,982] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:11,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:11,422] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:11,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:12,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:14,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:16,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:27:17,609] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 00:28:02,532 (deepspeed_trainer:228) INFO: 44epoch:train:3701-3800batch: iter_time=2.476, loss_ctc=60.341, loss_att=46.210, acc=0.720, loss=50.462, grad_norm=5.747, loss_scale=1.000, learning_rate=6.082e-05, step_time=0.433 [cnode7-018:0/16] 2024-12-11 00:28:39,518 (deepspeed_trainer:228) INFO: 44epoch:train:3801-3900batch: iter_time=1.060e-04, loss_ctc=64.938, loss_att=45.172, acc=0.746, loss=51.107, grad_norm=5.801, loss_scale=1.000, learning_rate=6.082e-05, step_time=0.370 [2024-12-11 00:29:16,535] [INFO] [logging.py:129:log_dist] [Rank 0] step=649000, skipped=0, lr=[np.float64(6.081111099422959e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:29:16,535] [INFO] [timer.py:264:stop] epoch=0/micro_step=49000/global_step=49000, RunningAvgSamplesPerSec=44.14161067834998, CurrSamplesPerSec=46.87151558197315, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:29:16,537 (deepspeed_trainer:228) INFO: 44epoch:train:3901-4000batch: iter_time=1.113e-04, loss_ctc=69.583, loss_att=52.940, acc=0.724, loss=57.947, grad_norm=6.499, loss_scale=1.000, learning_rate=6.081e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:29:53,688 (deepspeed_trainer:228) INFO: 44epoch:train:4001-4100batch: iter_time=1.121e-04, loss_ctc=74.060, loss_att=58.129, acc=0.718, loss=62.884, grad_norm=6.827, loss_scale=1.000, learning_rate=6.081e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:30:30,222 (deepspeed_trainer:228) INFO: 44epoch:train:4101-4200batch: iter_time=1.134e-04, loss_ctc=71.231, loss_att=46.841, acc=0.733, loss=54.171, grad_norm=6.464, loss_scale=1.000, learning_rate=6.080e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 00:31:06,936 (deepspeed_trainer:228) INFO: 44epoch:train:4201-4300batch: iter_time=1.106e-04, loss_ctc=66.298, loss_att=48.731, acc=0.736, loss=54.005, grad_norm=6.884, loss_scale=1.000, learning_rate=6.080e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:31:43,831 (deepspeed_trainer:228) INFO: 44epoch:train:4301-4400batch: iter_time=1.111e-04, loss_ctc=70.894, loss_att=51.949, acc=0.723, loss=57.628, grad_norm=6.666, loss_scale=1.000, learning_rate=6.079e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:32:20,652 (deepspeed_trainer:228) INFO: 44epoch:train:4401-4500batch: iter_time=1.108e-04, loss_ctc=64.109, loss_att=46.116, acc=0.729, loss=51.516, grad_norm=6.474, loss_scale=1.000, learning_rate=6.079e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 00:32:57,753 (deepspeed_trainer:228) INFO: 44epoch:train:4501-4600batch: iter_time=1.103e-04, loss_ctc=69.364, loss_att=47.641, acc=0.733, loss=54.146, grad_norm=6.607, loss_scale=1.000, learning_rate=6.079e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:33:34,300 (deepspeed_trainer:228) INFO: 44epoch:train:4601-4700batch: iter_time=1.104e-04, loss_ctc=63.803, loss_att=46.183, acc=0.721, loss=51.450, grad_norm=6.392, loss_scale=1.000, learning_rate=6.078e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:34:11,397 (deepspeed_trainer:228) INFO: 44epoch:train:4701-4800batch: iter_time=1.124e-04, loss_ctc=76.208, loss_att=55.175, acc=0.725, loss=61.493, grad_norm=6.645, loss_scale=1.000, learning_rate=6.078e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:34:47,724 (deepspeed_trainer:228) INFO: 44epoch:train:4801-4900batch: iter_time=1.087e-04, loss_ctc=59.646, loss_att=42.305, acc=0.737, loss=47.508, grad_norm=5.729, loss_scale=1.000, learning_rate=6.077e-05, step_time=0.363 [2024-12-11 00:35:24,364] [INFO] [logging.py:129:log_dist] [Rank 0] step=650000, skipped=0, lr=[np.float64(6.076431528325699e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:35:24,365] [INFO] [timer.py:264:stop] epoch=0/micro_step=50000/global_step=50000, RunningAvgSamplesPerSec=44.15090539888921, CurrSamplesPerSec=46.13998511808238, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:35:24,367 (deepspeed_trainer:228) INFO: 44epoch:train:4901-5000batch: iter_time=1.095e-04, loss_ctc=69.907, loss_att=54.482, acc=0.728, loss=59.107, grad_norm=6.608, loss_scale=1.000, learning_rate=6.077e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:36:01,401 (deepspeed_trainer:228) INFO: 44epoch:train:5001-5100batch: iter_time=1.099e-04, loss_ctc=72.213, loss_att=53.147, acc=0.725, loss=58.870, grad_norm=6.344, loss_scale=1.000, learning_rate=6.076e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:36:37,702 (deepspeed_trainer:228) INFO: 44epoch:train:5101-5200batch: iter_time=1.079e-04, loss_ctc=63.888, loss_att=43.953, acc=0.728, loss=49.950, grad_norm=6.018, loss_scale=1.000, learning_rate=6.076e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 00:37:14,322 (deepspeed_trainer:228) INFO: 44epoch:train:5201-5300batch: iter_time=1.090e-04, loss_ctc=70.352, loss_att=51.644, acc=0.719, loss=57.241, grad_norm=6.142, loss_scale=1.000, learning_rate=6.075e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:37:51,011 (deepspeed_trainer:228) INFO: 44epoch:train:5301-5400batch: iter_time=1.097e-04, loss_ctc=63.416, loss_att=46.937, acc=0.733, loss=51.874, grad_norm=5.527, loss_scale=1.000, learning_rate=6.075e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:38:28,009 (deepspeed_trainer:228) INFO: 44epoch:train:5401-5500batch: iter_time=1.120e-04, loss_ctc=77.077, loss_att=57.428, acc=0.730, loss=63.319, grad_norm=6.847, loss_scale=1.000, learning_rate=6.074e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:39:04,690 (deepspeed_trainer:228) INFO: 44epoch:train:5501-5600batch: iter_time=1.105e-04, loss_ctc=64.389, loss_att=52.559, acc=0.715, loss=56.088, grad_norm=6.175, loss_scale=1.000, learning_rate=6.074e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:39:17,423 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-018:0/16] 2024-12-11 00:39:43,085 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 00:39:59,643 (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: ) [cnode7-018:0/16] 2024-12-11 00:39:59,643 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 00:39:59,646 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 00:40:25,578] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:26,025] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:25,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:26,081] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:26,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:26,665] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:26,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:27,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:27,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,773] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:40:28,981] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:14,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:14,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:14,635] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:15,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:15,906] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:15,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:16,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:16,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:16,747] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:17,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:17,532] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:17,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:17,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:18,481] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:18,531] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:41:18,709] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:02,288] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:03,035] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:03,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:03,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:04,365] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:04,654] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:05,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:05,399] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:05,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:05,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:05,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:06,493] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:06,594] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:06,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:07,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:07,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:50,617] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:51,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:51,100] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:51,764] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:53,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:53,510] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:53,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:53,579] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:54,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:54,306] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:54,648] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:54,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:54,995] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:55,369] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:56,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:42:56,076] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 00:43:50,797 (deepspeed_trainer:228) INFO: 44epoch:train:5601-5700batch: iter_time=2.487, loss_ctc=64.233, loss_att=51.171, acc=0.730, loss=55.101, grad_norm=6.601, loss_scale=1.000, learning_rate=6.073e-05, step_time=0.379 [cnode7-018:0/16] 2024-12-11 00:44:28,153 (deepspeed_trainer:228) INFO: 44epoch:train:5701-5800batch: iter_time=1.083e-04, loss_ctc=63.845, loss_att=48.157, acc=0.743, loss=52.892, grad_norm=5.888, loss_scale=1.000, learning_rate=6.073e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 00:45:05,614 (deepspeed_trainer:228) INFO: 44epoch:train:5801-5900batch: iter_time=1.057e-04, loss_ctc=70.528, loss_att=51.435, acc=0.737, loss=57.190, grad_norm=6.595, loss_scale=1.000, learning_rate=6.072e-05, step_time=0.374 [2024-12-11 00:45:42,983] [INFO] [logging.py:129:log_dist] [Rank 0] step=651000, skipped=0, lr=[np.float64(6.0717627437787085e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:45:42,984] [INFO] [timer.py:264:stop] epoch=0/micro_step=51000/global_step=51000, RunningAvgSamplesPerSec=44.15398810040904, CurrSamplesPerSec=40.71284553687834, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:45:42,986 (deepspeed_trainer:228) INFO: 44epoch:train:5901-6000batch: iter_time=1.095e-04, loss_ctc=74.601, loss_att=61.108, acc=0.715, loss=65.172, grad_norm=7.221, loss_scale=1.000, learning_rate=6.072e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 00:46:19,897 (deepspeed_trainer:228) INFO: 44epoch:train:6001-6100batch: iter_time=1.063e-04, loss_ctc=72.155, loss_att=48.114, acc=0.738, loss=55.314, grad_norm=5.930, loss_scale=1.000, learning_rate=6.072e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:46:57,178 (deepspeed_trainer:228) INFO: 44epoch:train:6101-6200batch: iter_time=1.084e-04, loss_ctc=64.157, loss_att=47.648, acc=0.741, loss=52.602, grad_norm=5.838, loss_scale=1.000, learning_rate=6.071e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 00:47:34,039 (deepspeed_trainer:228) INFO: 44epoch:train:6201-6300batch: iter_time=1.073e-04, loss_ctc=75.276, loss_att=55.732, acc=0.726, loss=61.589, grad_norm=7.445, loss_scale=1.000, learning_rate=6.071e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 00:48:10,633 (deepspeed_trainer:228) INFO: 44epoch:train:6301-6400batch: iter_time=1.079e-04, loss_ctc=61.454, loss_att=42.945, acc=0.740, loss=48.499, grad_norm=5.657, loss_scale=1.000, learning_rate=6.070e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 00:48:47,620 (deepspeed_trainer:228) INFO: 44epoch:train:6401-6500batch: iter_time=1.097e-04, loss_ctc=68.906, loss_att=49.765, acc=0.733, loss=55.515, grad_norm=6.217, loss_scale=1.000, learning_rate=6.070e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:49:24,753 (deepspeed_trainer:228) INFO: 44epoch:train:6501-6600batch: iter_time=1.092e-04, loss_ctc=69.184, loss_att=51.728, acc=0.711, loss=56.953, grad_norm=6.923, loss_scale=1.000, learning_rate=6.069e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 00:50:01,725 (deepspeed_trainer:228) INFO: 44epoch:train:6601-6700batch: iter_time=1.104e-04, loss_ctc=72.051, loss_att=51.201, acc=0.739, loss=57.471, grad_norm=6.359, loss_scale=1.000, learning_rate=6.069e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:50:38,323 (deepspeed_trainer:228) INFO: 44epoch:train:6701-6800batch: iter_time=1.090e-04, loss_ctc=59.756, loss_att=44.463, acc=0.739, loss=49.062, grad_norm=5.694, loss_scale=1.000, learning_rate=6.068e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:51:14,982 (deepspeed_trainer:228) INFO: 44epoch:train:6801-6900batch: iter_time=1.100e-04, loss_ctc=68.125, loss_att=51.272, acc=0.739, loss=56.309, grad_norm=5.607, loss_scale=1.000, learning_rate=6.068e-05, step_time=0.366 [2024-12-11 00:51:51,665] [INFO] [logging.py:129:log_dist] [Rank 0] step=652000, skipped=0, lr=[np.float64(6.0671047044067e-05)], mom=[[0.9, 0.98]] [2024-12-11 00:51:51,666] [INFO] [timer.py:264:stop] epoch=0/micro_step=52000/global_step=52000, RunningAvgSamplesPerSec=44.161113917349944, CurrSamplesPerSec=45.446105504781315, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 00:51:51,667 (deepspeed_trainer:228) INFO: 44epoch:train:6901-7000batch: iter_time=1.115e-04, loss_ctc=69.398, loss_att=53.219, acc=0.722, loss=58.069, grad_norm=6.317, loss_scale=1.000, learning_rate=6.067e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:52:28,399 (deepspeed_trainer:228) INFO: 44epoch:train:7001-7100batch: iter_time=1.095e-04, loss_ctc=67.873, loss_att=47.515, acc=0.733, loss=53.626, grad_norm=6.131, loss_scale=1.000, learning_rate=6.067e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 00:53:06,077 (deepspeed_trainer:228) INFO: 44epoch:train:7101-7200batch: iter_time=1.105e-04, loss_ctc=75.086, loss_att=58.423, acc=0.723, loss=63.433, grad_norm=5.882, loss_scale=1.000, learning_rate=6.066e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-11 00:53:43,012 (deepspeed_trainer:228) INFO: 44epoch:train:7201-7300batch: iter_time=1.106e-04, loss_ctc=60.351, loss_att=42.582, acc=0.734, loss=47.919, grad_norm=5.878, loss_scale=1.000, learning_rate=6.066e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 00:54:19,911 (deepspeed_trainer:228) INFO: 44epoch:train:7301-7400batch: iter_time=1.073e-04, loss_ctc=72.987, loss_att=56.232, acc=0.742, loss=61.269, grad_norm=6.232, loss_scale=1.000, learning_rate=6.065e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 00:54:56,581 (deepspeed_trainer:228) INFO: 44epoch:train:7401-7500batch: iter_time=1.086e-04, loss_ctc=61.259, loss_att=50.560, acc=0.721, loss=53.781, grad_norm=5.856, loss_scale=1.000, learning_rate=6.065e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 00:54:59,664 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-018:0/16] 2024-12-11 00:55:27,026 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 00:55:43,474 (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: ) [cnode7-018:0/16] 2024-12-11 00:55:43,475 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 00:55:43,477 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 00:56:09,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:11,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:11,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:11,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:11,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:11,698] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,250] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,515] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,840] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:12,840] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:13,174] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:13,198] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:58,882] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:56:59,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:00,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:00,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:00,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,208] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,719] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:01,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:02,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:02,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:02,637] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:02,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:46,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:48,173] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:48,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:49,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:49,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:49,695] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:50,180] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:50,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:50,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:50,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:50,924] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:51,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:51,720] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:52,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:52,449] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:57:52,503] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:34,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:36,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:36,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:37,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:37,766] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:38,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:38,044] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:38,433] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:38,822] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:39,133] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:39,401] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:39,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:40,112] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:40,407] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:40,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 00:58:44,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 00:59:47,362 (deepspeed_trainer:228) INFO: 44epoch:train:7501-7600batch: iter_time=2.487, loss_ctc=61.331, loss_att=45.790, acc=0.743, loss=50.453, grad_norm=5.506, loss_scale=1.000, learning_rate=6.065e-05, step_time=0.420 [cnode7-018:0/16] 2024-12-11 01:00:24,140 (deepspeed_trainer:228) INFO: 44epoch:train:7601-7700batch: iter_time=1.091e-04, loss_ctc=66.802, loss_att=52.453, acc=0.731, loss=56.762, grad_norm=6.612, loss_scale=1.000, learning_rate=6.064e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:01:01,009 (deepspeed_trainer:228) INFO: 44epoch:train:7701-7800batch: iter_time=1.117e-04, loss_ctc=73.642, loss_att=53.121, acc=0.747, loss=59.268, grad_norm=7.278, loss_scale=1.000, learning_rate=6.064e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:01:37,680 (deepspeed_trainer:228) INFO: 44epoch:train:7801-7900batch: iter_time=1.126e-04, loss_ctc=68.606, loss_att=54.488, acc=0.719, loss=58.728, grad_norm=7.562, loss_scale=1.000, learning_rate=6.063e-05, step_time=0.366 [2024-12-11 01:02:14,518] [INFO] [logging.py:129:log_dist] [Rank 0] step=653000, skipped=0, lr=[np.float64(6.0624573690562375e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:02:14,518] [INFO] [timer.py:264:stop] epoch=0/micro_step=53000/global_step=53000, RunningAvgSamplesPerSec=44.15545506264029, CurrSamplesPerSec=45.646910350430005, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:02:14,520 (deepspeed_trainer:228) INFO: 44epoch:train:7901-8000batch: iter_time=1.138e-04, loss_ctc=73.387, loss_att=49.331, acc=0.740, loss=56.574, grad_norm=6.195, loss_scale=1.000, learning_rate=6.063e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:02:51,463 (deepspeed_trainer:228) INFO: 44epoch:train:8001-8100batch: iter_time=1.126e-04, loss_ctc=66.523, loss_att=50.164, acc=0.733, loss=55.068, grad_norm=6.487, loss_scale=1.000, learning_rate=6.062e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 01:03:28,214 (deepspeed_trainer:228) INFO: 44epoch:train:8101-8200batch: iter_time=1.124e-04, loss_ctc=73.022, loss_att=55.172, acc=0.725, loss=60.519, grad_norm=7.308, loss_scale=1.000, learning_rate=6.062e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:04:04,731 (deepspeed_trainer:228) INFO: 44epoch:train:8201-8300batch: iter_time=1.094e-04, loss_ctc=60.287, loss_att=43.250, acc=0.741, loss=48.375, grad_norm=5.813, loss_scale=1.000, learning_rate=6.061e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:04:41,300 (deepspeed_trainer:228) INFO: 44epoch:train:8301-8400batch: iter_time=1.128e-04, loss_ctc=66.760, loss_att=46.633, acc=0.732, loss=52.674, grad_norm=6.596, loss_scale=1.000, learning_rate=6.061e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:05:18,120 (deepspeed_trainer:228) INFO: 44epoch:train:8401-8500batch: iter_time=1.103e-04, loss_ctc=73.556, loss_att=55.884, acc=0.717, loss=61.174, grad_norm=6.304, loss_scale=1.000, learning_rate=6.060e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:05:54,560 (deepspeed_trainer:228) INFO: 44epoch:train:8501-8600batch: iter_time=1.117e-04, loss_ctc=67.101, loss_att=46.950, acc=0.738, loss=53.004, grad_norm=6.492, loss_scale=1.000, learning_rate=6.060e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 01:06:31,149 (deepspeed_trainer:228) INFO: 44epoch:train:8601-8700batch: iter_time=1.134e-04, loss_ctc=62.516, loss_att=47.778, acc=0.745, loss=52.202, grad_norm=5.873, loss_scale=1.000, learning_rate=6.059e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:07:07,841 (deepspeed_trainer:228) INFO: 44epoch:train:8701-8800batch: iter_time=1.112e-04, loss_ctc=64.118, loss_att=46.167, acc=0.743, loss=51.558, grad_norm=5.628, loss_scale=1.000, learning_rate=6.059e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:07:45,098 (deepspeed_trainer:228) INFO: 44epoch:train:8801-8900batch: iter_time=1.114e-04, loss_ctc=70.279, loss_att=53.543, acc=0.717, loss=58.562, grad_norm=7.231, loss_scale=1.000, learning_rate=6.059e-05, step_time=0.372 [2024-12-11 01:08:21,767] [INFO] [logging.py:129:log_dist] [Rank 0] step=654000, skipped=0, lr=[np.float64(6.057820696794208e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:08:21,768] [INFO] [timer.py:264:stop] epoch=0/micro_step=54000/global_step=54000, RunningAvgSamplesPerSec=44.165441413108745, CurrSamplesPerSec=46.888970869990345, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:08:21,770 (deepspeed_trainer:228) INFO: 44epoch:train:8901-9000batch: iter_time=1.120e-04, loss_ctc=71.304, loss_att=49.177, acc=0.731, loss=55.833, grad_norm=6.504, loss_scale=1.000, learning_rate=6.058e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:08:58,437 (deepspeed_trainer:228) INFO: 44epoch:train:9001-9100batch: iter_time=1.145e-04, loss_ctc=69.647, loss_att=55.353, acc=0.731, loss=59.674, grad_norm=6.130, loss_scale=1.000, learning_rate=6.058e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 01:09:34,953 (deepspeed_trainer:228) INFO: 44epoch:train:9101-9200batch: iter_time=1.080e-04, loss_ctc=65.024, loss_att=46.683, acc=0.734, loss=52.196, grad_norm=6.284, loss_scale=1.000, learning_rate=6.057e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:10:12,034 (deepspeed_trainer:228) INFO: 44epoch:train:9201-9300batch: iter_time=1.144e-04, loss_ctc=74.022, loss_att=57.665, acc=0.740, loss=62.571, grad_norm=6.086, loss_scale=1.000, learning_rate=6.057e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 01:10:42,769 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-018:0/16] 2024-12-11 01:11:10,161 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 01:11:26,824 (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: ) [cnode7-018:0/16] 2024-12-11 01:11:26,824 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 01:11:26,827 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 01:11:53,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:53,338] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:54,437] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:53,912] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:54,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,146] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:54,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:54,499] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,106] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:56,179] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,489] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:56,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:56,560] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:11:55,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:42,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:42,511] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:43,216] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:43,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:43,357] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:43,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:44,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:44,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:45,247] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:45,334] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:44,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:45,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:44,939] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:44,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:45,292] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:12:47,715] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:30,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:31,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:31,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:32,013] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:32,404] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:32,871] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:33,927] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:33,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:34,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:34,782] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:34,323] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:34,362] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:34,415] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:35,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:39,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:13:39,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:19,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:19,430] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:18,955] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:19,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:21,255] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:21,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:22,201] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:21,730] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:21,960] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:22,895] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:23,355] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:24,121] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:23,757] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:26,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:29,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:14:30,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 01:15:04,903 (deepspeed_trainer:228) INFO: 44epoch:train:9301-9400batch: iter_time=2.509, loss_ctc=59.105, loss_att=49.936, acc=0.717, loss=52.671, grad_norm=5.853, loss_scale=1.000, learning_rate=6.056e-05, step_time=0.419 [cnode7-018:0/16] 2024-12-11 01:15:41,971 (deepspeed_trainer:228) INFO: 44epoch:train:9401-9500batch: iter_time=1.050e-04, loss_ctc=63.316, loss_att=47.090, acc=0.738, loss=51.975, grad_norm=6.518, loss_scale=1.000, learning_rate=6.056e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 01:16:18,813 (deepspeed_trainer:228) INFO: 44epoch:train:9501-9600batch: iter_time=1.096e-04, loss_ctc=64.632, loss_att=50.542, acc=0.729, loss=54.775, grad_norm=6.028, loss_scale=1.000, learning_rate=6.055e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:16:55,724 (deepspeed_trainer:228) INFO: 44epoch:train:9601-9700batch: iter_time=1.075e-04, loss_ctc=74.427, loss_att=57.892, acc=0.728, loss=62.823, grad_norm=7.589, loss_scale=1.000, learning_rate=6.055e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 01:17:32,144 (deepspeed_trainer:228) INFO: 44epoch:train:9701-9800batch: iter_time=1.089e-04, loss_ctc=68.723, loss_att=48.623, acc=0.727, loss=54.641, grad_norm=6.299, loss_scale=1.000, learning_rate=6.054e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 01:18:08,861 (deepspeed_trainer:228) INFO: 44epoch:train:9801-9900batch: iter_time=1.100e-04, loss_ctc=70.906, loss_att=48.870, acc=0.734, loss=55.470, grad_norm=6.487, loss_scale=1.000, learning_rate=6.054e-05, step_time=0.367 [2024-12-11 01:18:45,410] [INFO] [logging.py:129:log_dist] [Rank 0] step=655000, skipped=0, lr=[np.float64(6.0531946469063104e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:18:45,411] [INFO] [timer.py:264:stop] epoch=0/micro_step=55000/global_step=55000, RunningAvgSamplesPerSec=44.163012024585086, CurrSamplesPerSec=46.348391000566494, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:18:45,413 (deepspeed_trainer:228) INFO: 44epoch:train:9901-10000batch: iter_time=1.119e-04, loss_ctc=68.458, loss_att=50.619, acc=0.724, loss=55.966, grad_norm=7.318, loss_scale=1.000, learning_rate=6.053e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:19:22,463 (deepspeed_trainer:228) INFO: 44epoch:train:10001-10100batch: iter_time=1.157e-04, loss_ctc=68.324, loss_att=52.177, acc=0.731, loss=57.003, grad_norm=6.554, loss_scale=1.000, learning_rate=6.053e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 01:19:59,036 (deepspeed_trainer:228) INFO: 44epoch:train:10101-10200batch: iter_time=1.161e-04, loss_ctc=61.159, loss_att=41.980, acc=0.739, loss=47.717, grad_norm=5.820, loss_scale=1.000, learning_rate=6.052e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:20:35,290 (deepspeed_trainer:228) INFO: 44epoch:train:10201-10300batch: iter_time=1.127e-04, loss_ctc=69.278, loss_att=49.432, acc=0.725, loss=55.394, grad_norm=6.447, loss_scale=1.000, learning_rate=6.052e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 01:21:12,158 (deepspeed_trainer:228) INFO: 44epoch:train:10301-10400batch: iter_time=1.148e-04, loss_ctc=74.336, loss_att=55.907, acc=0.712, loss=61.452, grad_norm=6.790, loss_scale=1.000, learning_rate=6.052e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:21:48,363 (deepspeed_trainer:228) INFO: 44epoch:train:10401-10500batch: iter_time=1.114e-04, loss_ctc=64.824, loss_att=42.843, acc=0.745, loss=49.423, grad_norm=5.771, loss_scale=1.000, learning_rate=6.051e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 01:22:24,582 (deepspeed_trainer:228) INFO: 44epoch:train:10501-10600batch: iter_time=1.083e-04, loss_ctc=62.390, loss_att=48.486, acc=0.732, loss=52.652, grad_norm=5.754, loss_scale=1.000, learning_rate=6.051e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 01:23:00,901 (deepspeed_trainer:228) INFO: 44epoch:train:10601-10700batch: iter_time=1.098e-04, loss_ctc=71.587, loss_att=53.439, acc=0.733, loss=58.896, grad_norm=6.764, loss_scale=1.000, learning_rate=6.050e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:23:37,519 (deepspeed_trainer:228) INFO: 44epoch:train:10701-10800batch: iter_time=1.102e-04, loss_ctc=64.389, loss_att=45.875, acc=0.728, loss=51.416, grad_norm=6.282, loss_scale=1.000, learning_rate=6.050e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:24:14,469 (deepspeed_trainer:228) INFO: 44epoch:train:10801-10900batch: iter_time=1.109e-04, loss_ctc=68.542, loss_att=48.887, acc=0.719, loss=54.798, grad_norm=6.854, loss_scale=1.000, learning_rate=6.049e-05, step_time=0.370 [2024-12-11 01:24:51,398] [INFO] [logging.py:129:log_dist] [Rank 0] step=656000, skipped=0, lr=[np.float64(6.048579178895551e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:24:51,399] [INFO] [timer.py:264:stop] epoch=0/micro_step=56000/global_step=56000, RunningAvgSamplesPerSec=44.17571234610683, CurrSamplesPerSec=46.487314632794394, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:24:51,400 (deepspeed_trainer:228) INFO: 44epoch:train:10901-11000batch: iter_time=1.154e-04, loss_ctc=68.172, loss_att=52.940, acc=0.732, loss=57.499, grad_norm=6.001, loss_scale=1.000, learning_rate=6.049e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 01:25:28,589 (deepspeed_trainer:228) INFO: 44epoch:train:11001-11100batch: iter_time=1.136e-04, loss_ctc=74.422, loss_att=53.984, acc=0.731, loss=60.111, grad_norm=6.714, loss_scale=1.000, learning_rate=6.048e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 01:26:05,542 (deepspeed_trainer:228) INFO: 44epoch:train:11101-11200batch: iter_time=1.136e-04, loss_ctc=68.257, loss_att=56.347, acc=0.720, loss=59.918, grad_norm=6.220, loss_scale=1.000, learning_rate=6.048e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 01:26:26,905 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-018:0/16] 2024-12-11 01:26:54,744 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 01:27:11,330 (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: ) [cnode7-018:0/16] 2024-12-11 01:27:11,330 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 01:27:11,332 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 01:27:37,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:38,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:37,865] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:38,846] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:38,339] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:38,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,501] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,556] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:38,950] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,599] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:40,477] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:39,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:27:40,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:26,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:26,756] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:28,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:27,565] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:27,577] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:27,804] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:28,656] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:28,257] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:29,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:29,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:28,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:29,817] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:30,908] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:32,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:33,018] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:28:32,909] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:15,105] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:15,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:16,328] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:16,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:16,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,718] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:18,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:17,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:23,149] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:24,120] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:24,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:29:24,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:02,864] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:03,321] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:04,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:05,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:04,781] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:05,666] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:05,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:06,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:05,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:06,425] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:05,823] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:06,128] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:13,801] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:13,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:14,000] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:30:14,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 01:30:58,693 (deepspeed_trainer:228) INFO: 44epoch:train:11201-11300batch: iter_time=2.489, loss_ctc=59.716, loss_att=46.721, acc=0.722, loss=50.624, grad_norm=6.251, loss_scale=1.000, learning_rate=6.047e-05, step_time=0.442 [cnode7-018:0/16] 2024-12-11 01:31:35,458 (deepspeed_trainer:228) INFO: 44epoch:train:11301-11400batch: iter_time=1.055e-04, loss_ctc=64.229, loss_att=44.492, acc=0.749, loss=50.399, grad_norm=5.984, loss_scale=1.000, learning_rate=6.047e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:32:12,170 (deepspeed_trainer:228) INFO: 44epoch:train:11401-11500batch: iter_time=1.080e-04, loss_ctc=69.029, loss_att=52.657, acc=0.726, loss=57.580, grad_norm=5.843, loss_scale=1.000, learning_rate=6.047e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:32:48,925 (deepspeed_trainer:228) INFO: 44epoch:train:11501-11600batch: iter_time=1.086e-04, loss_ctc=73.400, loss_att=58.273, acc=0.718, loss=62.795, grad_norm=7.004, loss_scale=1.000, learning_rate=6.046e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:33:25,712 (deepspeed_trainer:228) INFO: 44epoch:train:11601-11700batch: iter_time=1.117e-04, loss_ctc=70.465, loss_att=46.569, acc=0.736, loss=53.723, grad_norm=6.149, loss_scale=1.000, learning_rate=6.046e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:34:02,065 (deepspeed_trainer:228) INFO: 44epoch:train:11701-11800batch: iter_time=1.100e-04, loss_ctc=65.549, loss_att=48.437, acc=0.739, loss=53.570, grad_norm=6.277, loss_scale=1.000, learning_rate=6.045e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:34:38,337 (deepspeed_trainer:228) INFO: 44epoch:train:11801-11900batch: iter_time=1.078e-04, loss_ctc=69.817, loss_att=50.969, acc=0.727, loss=56.642, grad_norm=6.900, loss_scale=1.000, learning_rate=6.045e-05, step_time=0.362 [2024-12-11 01:35:14,522] [INFO] [logging.py:129:log_dist] [Rank 0] step=657000, skipped=0, lr=[np.float64(6.043974252480754e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:35:14,522] [INFO] [timer.py:264:stop] epoch=0/micro_step=57000/global_step=57000, RunningAvgSamplesPerSec=44.169825590371936, CurrSamplesPerSec=46.621955516412484, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:35:14,524 (deepspeed_trainer:228) INFO: 44epoch:train:11901-12000batch: iter_time=1.097e-04, loss_ctc=63.678, loss_att=46.159, acc=0.732, loss=51.393, grad_norm=6.622, loss_scale=1.000, learning_rate=6.044e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 01:35:50,907 (deepspeed_trainer:228) INFO: 44epoch:train:12001-12100batch: iter_time=1.090e-04, loss_ctc=68.661, loss_att=47.483, acc=0.737, loss=53.839, grad_norm=6.708, loss_scale=1.000, learning_rate=6.044e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:36:27,471 (deepspeed_trainer:228) INFO: 44epoch:train:12101-12200batch: iter_time=1.079e-04, loss_ctc=63.627, loss_att=45.595, acc=0.723, loss=50.993, grad_norm=6.038, loss_scale=1.000, learning_rate=6.043e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 01:37:03,876 (deepspeed_trainer:228) INFO: 44epoch:train:12201-12300batch: iter_time=1.083e-04, loss_ctc=75.421, loss_att=53.961, acc=0.728, loss=60.388, grad_norm=6.871, loss_scale=1.000, learning_rate=6.043e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 01:37:40,190 (deepspeed_trainer:228) INFO: 44epoch:train:12301-12400batch: iter_time=1.096e-04, loss_ctc=58.799, loss_att=41.532, acc=0.740, loss=46.726, grad_norm=5.990, loss_scale=1.000, learning_rate=6.042e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:38:17,237 (deepspeed_trainer:228) INFO: 44epoch:train:12401-12500batch: iter_time=1.081e-04, loss_ctc=70.002, loss_att=54.461, acc=0.730, loss=59.125, grad_norm=6.486, loss_scale=1.000, learning_rate=6.042e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 01:38:54,010 (deepspeed_trainer:228) INFO: 44epoch:train:12501-12600batch: iter_time=1.105e-04, loss_ctc=71.581, loss_att=53.107, acc=0.727, loss=58.658, grad_norm=6.331, loss_scale=1.000, learning_rate=6.041e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:39:30,348 (deepspeed_trainer:228) INFO: 44epoch:train:12601-12700batch: iter_time=1.106e-04, loss_ctc=62.506, loss_att=42.922, acc=0.735, loss=48.801, grad_norm=6.410, loss_scale=1.000, learning_rate=6.041e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:40:06,705 (deepspeed_trainer:228) INFO: 44epoch:train:12701-12800batch: iter_time=1.073e-04, loss_ctc=69.459, loss_att=51.364, acc=0.722, loss=56.795, grad_norm=6.371, loss_scale=1.000, learning_rate=6.041e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:40:43,129 (deepspeed_trainer:228) INFO: 44epoch:train:12801-12900batch: iter_time=1.074e-04, loss_ctc=62.776, loss_att=47.078, acc=0.734, loss=51.764, grad_norm=5.522, loss_scale=1.000, learning_rate=6.040e-05, step_time=0.364 [2024-12-11 01:41:19,856] [INFO] [logging.py:129:log_dist] [Rank 0] step=658000, skipped=0, lr=[np.float64(6.039379827595087e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:41:19,857] [INFO] [timer.py:264:stop] epoch=0/micro_step=58000/global_step=58000, RunningAvgSamplesPerSec=44.18276521059398, CurrSamplesPerSec=45.432045182585576, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:41:19,859 (deepspeed_trainer:228) INFO: 44epoch:train:12901-13000batch: iter_time=1.095e-04, loss_ctc=76.327, loss_att=56.270, acc=0.734, loss=62.248, grad_norm=6.383, loss_scale=1.000, learning_rate=6.040e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:41:55,288 (deepspeed_trainer:228) INFO: 44epoch:train:13001-13100batch: iter_time=1.116e-04, loss_ctc=63.354, loss_att=52.993, acc=0.715, loss=56.091, grad_norm=5.724, loss_scale=1.000, learning_rate=6.039e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:42:07,475 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-018:0/16] 2024-12-11 01:42:34,967 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 01:42:51,083 (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: ) [cnode7-018:0/16] 2024-12-11 01:42:51,083 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 01:42:51,085 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 01:43:18,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:18,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:20,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:19,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:21,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:21,458] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:19,664] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:19,692] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:21,601] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:21,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:20,102] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:21,991] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:22,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:22,192] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:20,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:43:20,862] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:09,741] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:07,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:10,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:08,455] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:10,625] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:08,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:10,965] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:11,015] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:09,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:09,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:09,802] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:10,634] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:11,325] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:13,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:14,367] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:14,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:56,826] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:58,867] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:58,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:57,150] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:57,558] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:59,530] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:59,574] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:59,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:58,277] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:58,613] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:59,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:44:59,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:01,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:04,094] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:05,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:06,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:44,495] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:45,222] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:47,393] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:45,580] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:47,600] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:48,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:46,691] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:47,669] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:47,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:50,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:51,067] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:51,387] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:53,411] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:52,423] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:55,204] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:45:56,813] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 01:46:49,727 (deepspeed_trainer:228) INFO: 44epoch:train:13101-13200batch: iter_time=2.481, loss_ctc=63.775, loss_att=50.962, acc=0.733, loss=54.827, grad_norm=6.356, loss_scale=1.000, learning_rate=6.039e-05, step_time=0.463 [cnode7-018:0/16] 2024-12-11 01:47:27,306 (deepspeed_trainer:228) INFO: 44epoch:train:13201-13300batch: iter_time=1.029e-04, loss_ctc=62.837, loss_att=47.123, acc=0.746, loss=51.816, grad_norm=6.473, loss_scale=1.000, learning_rate=6.038e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-11 01:48:04,883 (deepspeed_trainer:228) INFO: 44epoch:train:13301-13400batch: iter_time=1.058e-04, loss_ctc=69.634, loss_att=51.155, acc=0.739, loss=56.672, grad_norm=6.113, loss_scale=1.000, learning_rate=6.038e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-11 01:48:42,693 (deepspeed_trainer:228) INFO: 44epoch:train:13401-13500batch: iter_time=1.125e-04, loss_ctc=72.896, loss_att=60.846, acc=0.717, loss=64.468, grad_norm=7.371, loss_scale=1.000, learning_rate=6.037e-05, step_time=0.377 [cnode7-018:0/16] 2024-12-11 01:49:19,347 (deepspeed_trainer:228) INFO: 44epoch:train:13501-13600batch: iter_time=1.105e-04, loss_ctc=71.876, loss_att=47.830, acc=0.740, loss=55.049, grad_norm=6.424, loss_scale=1.000, learning_rate=6.037e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:49:56,015 (deepspeed_trainer:228) INFO: 44epoch:train:13601-13700batch: iter_time=1.138e-04, loss_ctc=63.299, loss_att=47.239, acc=0.742, loss=52.038, grad_norm=5.814, loss_scale=1.000, learning_rate=6.036e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:50:32,940 (deepspeed_trainer:228) INFO: 44epoch:train:13701-13800batch: iter_time=1.138e-04, loss_ctc=73.360, loss_att=54.968, acc=0.729, loss=60.472, grad_norm=9.345, loss_scale=1.000, learning_rate=6.036e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 01:51:09,395 (deepspeed_trainer:228) INFO: 44epoch:train:13801-13900batch: iter_time=1.086e-04, loss_ctc=61.444, loss_att=43.140, acc=0.741, loss=48.624, grad_norm=5.905, loss_scale=1.000, learning_rate=6.035e-05, step_time=0.364 [2024-12-11 01:51:45,893] [INFO] [logging.py:129:log_dist] [Rank 0] step=659000, skipped=0, lr=[np.float64(6.0347958643845976e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:51:45,894] [INFO] [timer.py:264:stop] epoch=0/micro_step=59000/global_step=59000, RunningAvgSamplesPerSec=44.16869571902852, CurrSamplesPerSec=47.0710418399391, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:51:45,895 (deepspeed_trainer:228) INFO: 44epoch:train:13901-14000batch: iter_time=1.097e-04, loss_ctc=67.497, loss_att=48.923, acc=0.737, loss=54.489, grad_norm=6.775, loss_scale=1.000, learning_rate=6.035e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:52:22,211 (deepspeed_trainer:228) INFO: 44epoch:train:14001-14100batch: iter_time=1.120e-04, loss_ctc=68.419, loss_att=51.337, acc=0.714, loss=56.467, grad_norm=6.733, loss_scale=1.000, learning_rate=6.035e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 01:52:58,715 (deepspeed_trainer:228) INFO: 44epoch:train:14101-14200batch: iter_time=1.143e-04, loss_ctc=71.935, loss_att=51.117, acc=0.739, loss=57.362, grad_norm=6.004, loss_scale=1.000, learning_rate=6.034e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:53:35,451 (deepspeed_trainer:228) INFO: 44epoch:train:14201-14300batch: iter_time=1.118e-04, loss_ctc=59.563, loss_att=43.766, acc=0.742, loss=48.501, grad_norm=6.117, loss_scale=1.000, learning_rate=6.034e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:54:12,046 (deepspeed_trainer:228) INFO: 44epoch:train:14301-14400batch: iter_time=1.107e-04, loss_ctc=67.536, loss_att=51.059, acc=0.740, loss=55.988, grad_norm=5.731, loss_scale=1.000, learning_rate=6.033e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:54:48,607 (deepspeed_trainer:228) INFO: 44epoch:train:14401-14500batch: iter_time=1.154e-04, loss_ctc=69.231, loss_att=52.636, acc=0.725, loss=57.608, grad_norm=6.161, loss_scale=1.000, learning_rate=6.033e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 01:55:25,416 (deepspeed_trainer:228) INFO: 44epoch:train:14501-14600batch: iter_time=1.121e-04, loss_ctc=66.894, loss_att=46.987, acc=0.735, loss=52.975, grad_norm=6.422, loss_scale=1.000, learning_rate=6.032e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 01:56:02,587 (deepspeed_trainer:228) INFO: 44epoch:train:14601-14700batch: iter_time=1.095e-04, loss_ctc=74.427, loss_att=57.834, acc=0.725, loss=62.798, grad_norm=6.574, loss_scale=1.000, learning_rate=6.032e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 01:56:39,278 (deepspeed_trainer:228) INFO: 44epoch:train:14701-14800batch: iter_time=1.123e-04, loss_ctc=59.661, loss_att=42.297, acc=0.736, loss=47.521, grad_norm=5.836, loss_scale=1.000, learning_rate=6.031e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 01:57:16,496 (deepspeed_trainer:228) INFO: 44epoch:train:14801-14900batch: iter_time=1.155e-04, loss_ctc=72.211, loss_att=55.835, acc=0.743, loss=60.774, grad_norm=5.973, loss_scale=1.000, learning_rate=6.031e-05, step_time=0.372 [2024-12-11 01:57:53,236] [INFO] [logging.py:129:log_dist] [Rank 0] step=660000, skipped=0, lr=[np.float64(6.0302223232067576e-05)], mom=[[0.9, 0.98]] [2024-12-11 01:57:53,236] [INFO] [timer.py:264:stop] epoch=0/micro_step=60000/global_step=60000, RunningAvgSamplesPerSec=44.17726323860428, CurrSamplesPerSec=47.522166247747606, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 01:57:53,238 (deepspeed_trainer:228) INFO: 44epoch:train:14901-15000batch: iter_time=1.084e-04, loss_ctc=61.341, loss_att=50.559, acc=0.723, loss=53.813, grad_norm=6.098, loss_scale=1.000, learning_rate=6.030e-05, step_time=0.367 [2024-12-11 01:58:07,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:09,913] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:09,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,620] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:10,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,752] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:10,270] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:10,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:10,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,894] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,915] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:08,967] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:09,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:22,172] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:24,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:25,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:25,923] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:24,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:25,996] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:25,061] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:25,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:26,970] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:27,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:27,219] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:27,231] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:26,097] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:26,168] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:26,262] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:26,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:37,299] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:40,512] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:41,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:40,027] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:41,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:41,896] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:40,614] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:40,902] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:43,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:43,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:41,755] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:43,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:43,642] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:42,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:42,874] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:42,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:52,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:55,687] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:57,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:55,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:56,267] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:56,611] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:58,156] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:58,506] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:57,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:58,686] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:59,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:59,269] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:58,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:59,937] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:59,316] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:58:59,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 01:59:11,750] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 44 is about to be saved! [2024-12-11 01:59:11,781] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/mp_rank_00_model_states.pt [2024-12-11 01:59:11,781] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/mp_rank_00_model_states.pt... [2024-12-11 01:59:13,658] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/mp_rank_00_model_states.pt. [2024-12-11 01:59:15,264] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,265] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,815] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,816] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,266] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,817] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,817] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,267] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,818] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,268] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,819] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,820] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,271] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,272] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-11 01:59:13,822] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,273] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-11 01:59:15,845] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-11 01:59:15,846] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-12-11 01:59:15,846] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,453] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,453] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,453] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:16,019] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,019] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-12-11 01:59:16,019] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:16,019] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,020] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-12-11 01:59:16,020] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,576] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,576] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,576] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,585] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,055] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,055] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-12-11 01:59:16,056] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,606] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,606] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,606] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:16,056] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,056] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-12-11 01:59:16,056] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:16,069] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-11 01:59:16,070] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-12-11 01:59:16,070] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,622] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,622] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,657] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,657] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,657] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,660] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,660] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,660] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,667] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,668] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,668] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:14,676] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-11 01:59:14,676] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-12-11 01:59:14,676] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:17,238] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-11 01:59:17,238] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-12-11 01:59:17,238] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [2024-12-11 01:59:17,462] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-11 01:59:17,462] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_44/44/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-12-11 01:59:17,462] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 44 is ready now! [cnode7-018:0/16] 2024-12-11 01:59:16,023 (deepspeed_trainer:158) INFO: 44epoch results: [train] iter_time=0.132, loss_ctc=68.112, loss_att=50.390, acc=0.730, loss=55.707, grad_norm=6.421, loss_scale=1.000, learning_rate=6.065e-05, step_time=0.371, time=2 hours, 6 minutes and 2.52 seconds, total_count=660044, gpu_max_cached_mem_GB=28.656, [valid] loss_ctc=4.031, cer_ctc=0.099, loss_att=9.438, acc=0.804, cer=0.389, wer=1.000, loss=7.812, time=1 minute and 10.37 seconds, total_count=44, gpu_max_cached_mem_GB=28.656 [cnode7-018:0/16] 2024-12-11 01:59:17,524 (multiple_iter_factory:32) INFO: Building 0th iter-factory... [cnode7-018:0/16] 2024-12-11 01:59:44,136 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 01:59:59,624 (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: ) [cnode7-018:0/16] 2024-12-11 01:59:59,624 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.5, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 01:59:59,627 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 02:00:22,780] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:22,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:23,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:23,524] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:23,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:22,261] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:24,214] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:24,494] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:23,079] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:23,137] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:25,191] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:24,011] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:24,080] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:24,561] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:25,096] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:00:25,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:10,975] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:11,370] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:11,432] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:12,014] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:12,079] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:10,733] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:11,189] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:12,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:12,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:11,667] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:11,763] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:13,837] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:12,989] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:13,303] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:13,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:15,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:58,521] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:58,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,206] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,382] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:00,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,449] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:01:59,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:01,285] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:00,724] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:01,452] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:01,670] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:04,684] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:04,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:45,731] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:46,995] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,087] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,356] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:46,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,019] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:48,782] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:49,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:47,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:48,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:50,020] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:52,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:54,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:02:55,651] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 02:03:58,617 (deepspeed_trainer:228) INFO: 45epoch:train:1-100batch: iter_time=2.361, loss_ctc=74.551, loss_att=56.045, acc=0.721, loss=61.627, grad_norm=6.200, loss_scale=1.000, learning_rate=6.030e-05, step_time=0.450 [cnode7-018:0/16] 2024-12-11 02:04:35,669 (deepspeed_trainer:228) INFO: 45epoch:train:101-200batch: iter_time=1.070e-04, loss_ctc=66.864, loss_att=51.203, acc=0.729, loss=55.892, grad_norm=5.903, loss_scale=1.000, learning_rate=6.030e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:05:12,377 (deepspeed_trainer:228) INFO: 45epoch:train:201-300batch: iter_time=1.158e-04, loss_ctc=60.071, loss_att=45.013, acc=0.736, loss=49.518, grad_norm=5.966, loss_scale=1.000, learning_rate=6.029e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:05:49,005 (deepspeed_trainer:228) INFO: 45epoch:train:301-400batch: iter_time=1.091e-04, loss_ctc=64.326, loss_att=45.499, acc=0.732, loss=51.130, grad_norm=5.846, loss_scale=1.000, learning_rate=6.029e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:06:25,840 (deepspeed_trainer:228) INFO: 45epoch:train:401-500batch: iter_time=1.107e-04, loss_ctc=73.538, loss_att=53.473, acc=0.711, loss=59.480, grad_norm=7.121, loss_scale=1.000, learning_rate=6.028e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:07:02,356 (deepspeed_trainer:228) INFO: 45epoch:train:501-600batch: iter_time=1.087e-04, loss_ctc=63.557, loss_att=43.902, acc=0.727, loss=49.809, grad_norm=5.721, loss_scale=1.000, learning_rate=6.028e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 02:07:39,492 (deepspeed_trainer:228) INFO: 45epoch:train:601-700batch: iter_time=1.115e-04, loss_ctc=78.913, loss_att=64.812, acc=0.709, loss=69.044, grad_norm=7.277, loss_scale=1.000, learning_rate=6.027e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 02:08:16,220 (deepspeed_trainer:228) INFO: 45epoch:train:701-800batch: iter_time=1.129e-04, loss_ctc=67.708, loss_att=46.149, acc=0.743, loss=52.614, grad_norm=7.236, loss_scale=1.000, learning_rate=6.027e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:08:53,083 (deepspeed_trainer:228) INFO: 45epoch:train:801-900batch: iter_time=1.092e-04, loss_ctc=65.400, loss_att=53.053, acc=0.714, loss=56.737, grad_norm=6.706, loss_scale=1.000, learning_rate=6.026e-05, step_time=0.368 [2024-12-11 02:09:30,236] [INFO] [logging.py:129:log_dist] [Rank 0] step=661000, skipped=0, lr=[np.float64(6.0256591646290296e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:09:30,237] [INFO] [timer.py:264:stop] epoch=0/micro_step=61000/global_step=61000, RunningAvgSamplesPerSec=44.166857778020834, CurrSamplesPerSec=46.03092585815222, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:09:30,238 (deepspeed_trainer:228) INFO: 45epoch:train:901-1000batch: iter_time=1.100e-04, loss_ctc=66.809, loss_att=54.032, acc=0.727, loss=57.877, grad_norm=6.728, loss_scale=1.000, learning_rate=6.026e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 02:10:07,919 (deepspeed_trainer:228) INFO: 45epoch:train:1001-1100batch: iter_time=1.096e-04, loss_ctc=76.052, loss_att=64.272, acc=0.709, loss=67.754, grad_norm=6.834, loss_scale=1.000, learning_rate=6.025e-05, step_time=0.376 [cnode7-018:0/16] 2024-12-11 02:10:45,203 (deepspeed_trainer:228) INFO: 45epoch:train:1101-1200batch: iter_time=1.086e-04, loss_ctc=58.054, loss_att=45.246, acc=0.723, loss=49.102, grad_norm=6.218, loss_scale=1.000, learning_rate=6.025e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 02:11:22,691 (deepspeed_trainer:228) INFO: 45epoch:train:1201-1300batch: iter_time=1.111e-04, loss_ctc=67.949, loss_att=52.812, acc=0.725, loss=57.352, grad_norm=6.338, loss_scale=1.000, learning_rate=6.025e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 02:11:59,938 (deepspeed_trainer:228) INFO: 45epoch:train:1301-1400batch: iter_time=1.128e-04, loss_ctc=65.356, loss_att=47.744, acc=0.725, loss=53.014, grad_norm=6.283, loss_scale=1.000, learning_rate=6.024e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 02:12:36,790 (deepspeed_trainer:228) INFO: 45epoch:train:1401-1500batch: iter_time=1.101e-04, loss_ctc=71.942, loss_att=55.013, acc=0.734, loss=60.086, grad_norm=6.531, loss_scale=1.000, learning_rate=6.024e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:13:13,732 (deepspeed_trainer:228) INFO: 45epoch:train:1501-1600batch: iter_time=1.087e-04, loss_ctc=71.642, loss_att=49.938, acc=0.718, loss=56.431, grad_norm=7.171, loss_scale=1.000, learning_rate=6.023e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 02:13:50,343 (deepspeed_trainer:228) INFO: 45epoch:train:1601-1700batch: iter_time=1.113e-04, loss_ctc=68.817, loss_att=52.836, acc=0.731, loss=57.644, grad_norm=6.006, loss_scale=1.000, learning_rate=6.023e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:14:27,256 (deepspeed_trainer:228) INFO: 45epoch:train:1701-1800batch: iter_time=1.113e-04, loss_ctc=69.581, loss_att=57.246, acc=0.727, loss=60.955, grad_norm=5.939, loss_scale=1.000, learning_rate=6.022e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 02:14:58,739 (multiple_iter_factory:32) INFO: Building 1th iter-factory... [cnode7-018:0/16] 2024-12-11 02:15:25,832 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 02:15:42,863 (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: ) [cnode7-018:0/16] 2024-12-11 02:15:42,863 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.0, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 02:15:42,865 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 02:16:12,039] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:12,119] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:13,675] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:13,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:12,535] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,167] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:12,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:12,816] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:12,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,315] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,345] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,347] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:14,398] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:13,005] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:16:13,054] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:01,266] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:03,291] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:02,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:02,336] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:02,591] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:02,657] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:02,808] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:03,067] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:04,771] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:04,806] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:04,853] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:04,888] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:05,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:05,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:05,233] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:05,340] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:51,110] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:49,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:50,605] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:50,968] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:51,380] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:52,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:51,770] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:53,325] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:51,910] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:54,297] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:54,379] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:54,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:54,825] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:53,954] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:57,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:17:56,359] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:39,005] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:37,546] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:37,905] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:38,689] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:39,045] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:41,701] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:41,744] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:42,468] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:40,872] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:43,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:43,628] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:44,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:43,068] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:44,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:48,900] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:18:47,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 02:19:23,015 (deepspeed_trainer:228) INFO: 45epoch:train:1801-1900batch: iter_time=2.496, loss_ctc=66.852, loss_att=51.424, acc=0.731, loss=56.047, grad_norm=6.301, loss_scale=1.000, learning_rate=6.022e-05, step_time=0.467 [2024-12-11 02:20:00,468] [INFO] [logging.py:129:log_dist] [Rank 0] step=662000, skipped=0, lr=[np.float64(6.021106349427429e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:20:00,469] [INFO] [timer.py:264:stop] epoch=0/micro_step=62000/global_step=62000, RunningAvgSamplesPerSec=44.14777269752191, CurrSamplesPerSec=45.345414291045785, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:20:00,470 (deepspeed_trainer:228) INFO: 45epoch:train:1901-2000batch: iter_time=1.044e-04, loss_ctc=71.087, loss_att=52.868, acc=0.725, loss=58.346, grad_norm=6.164, loss_scale=1.000, learning_rate=6.021e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 02:20:37,265 (deepspeed_trainer:228) INFO: 45epoch:train:2001-2100batch: iter_time=1.057e-04, loss_ctc=65.204, loss_att=50.773, acc=0.737, loss=55.137, grad_norm=5.816, loss_scale=1.000, learning_rate=6.021e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:21:13,931 (deepspeed_trainer:228) INFO: 45epoch:train:2101-2200batch: iter_time=1.097e-04, loss_ctc=60.712, loss_att=44.683, acc=0.740, loss=49.474, grad_norm=5.464, loss_scale=1.000, learning_rate=6.020e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:21:50,703 (deepspeed_trainer:228) INFO: 45epoch:train:2201-2300batch: iter_time=1.101e-04, loss_ctc=69.772, loss_att=49.161, acc=0.729, loss=55.338, grad_norm=6.240, loss_scale=1.000, learning_rate=6.020e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:22:27,353 (deepspeed_trainer:228) INFO: 45epoch:train:2301-2400batch: iter_time=1.064e-04, loss_ctc=63.432, loss_att=48.453, acc=0.726, loss=52.931, grad_norm=6.807, loss_scale=1.000, learning_rate=6.020e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:23:04,358 (deepspeed_trainer:228) INFO: 45epoch:train:2401-2500batch: iter_time=1.067e-04, loss_ctc=67.396, loss_att=51.528, acc=0.716, loss=56.275, grad_norm=6.279, loss_scale=1.000, learning_rate=6.019e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 02:23:41,700 (deepspeed_trainer:228) INFO: 45epoch:train:2501-2600batch: iter_time=1.067e-04, loss_ctc=78.518, loss_att=61.494, acc=0.731, loss=66.579, grad_norm=6.373, loss_scale=1.000, learning_rate=6.019e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 02:24:18,333 (deepspeed_trainer:228) INFO: 45epoch:train:2601-2700batch: iter_time=1.066e-04, loss_ctc=63.782, loss_att=45.855, acc=0.739, loss=51.233, grad_norm=6.243, loss_scale=1.000, learning_rate=6.018e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:24:54,945 (deepspeed_trainer:228) INFO: 45epoch:train:2701-2800batch: iter_time=1.084e-04, loss_ctc=65.102, loss_att=56.818, acc=0.717, loss=59.328, grad_norm=6.531, loss_scale=1.000, learning_rate=6.018e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:25:31,466 (deepspeed_trainer:228) INFO: 45epoch:train:2801-2900batch: iter_time=1.091e-04, loss_ctc=72.896, loss_att=57.416, acc=0.722, loss=62.071, grad_norm=7.022, loss_scale=1.000, learning_rate=6.017e-05, step_time=0.365 [2024-12-11 02:26:08,278] [INFO] [logging.py:129:log_dist] [Rank 0] step=663000, skipped=0, lr=[np.float64(6.01656383858512e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:26:08,279] [INFO] [timer.py:264:stop] epoch=0/micro_step=63000/global_step=63000, RunningAvgSamplesPerSec=44.15484136804342, CurrSamplesPerSec=45.57412433830509, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:26:08,281 (deepspeed_trainer:228) INFO: 45epoch:train:2901-3000batch: iter_time=1.080e-04, loss_ctc=67.234, loss_att=57.844, acc=0.728, loss=60.672, grad_norm=5.961, loss_scale=1.000, learning_rate=6.017e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:26:45,182 (deepspeed_trainer:228) INFO: 45epoch:train:3001-3100batch: iter_time=1.092e-04, loss_ctc=60.560, loss_att=50.179, acc=0.728, loss=53.275, grad_norm=5.960, loss_scale=1.000, learning_rate=6.016e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:27:21,860 (deepspeed_trainer:228) INFO: 45epoch:train:3101-3200batch: iter_time=1.107e-04, loss_ctc=64.388, loss_att=47.249, acc=0.737, loss=52.385, grad_norm=5.752, loss_scale=1.000, learning_rate=6.016e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:27:59,161 (deepspeed_trainer:228) INFO: 45epoch:train:3201-3300batch: iter_time=1.106e-04, loss_ctc=68.919, loss_att=50.397, acc=0.736, loss=55.971, grad_norm=5.902, loss_scale=1.000, learning_rate=6.015e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 02:28:36,245 (deepspeed_trainer:228) INFO: 45epoch:train:3301-3400batch: iter_time=1.086e-04, loss_ctc=69.851, loss_att=53.046, acc=0.724, loss=58.099, grad_norm=6.224, loss_scale=1.000, learning_rate=6.015e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:29:13,653 (deepspeed_trainer:228) INFO: 45epoch:train:3401-3500batch: iter_time=1.075e-04, loss_ctc=68.394, loss_att=48.809, acc=0.737, loss=54.711, grad_norm=6.088, loss_scale=1.000, learning_rate=6.015e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 02:29:50,685 (deepspeed_trainer:228) INFO: 45epoch:train:3501-3600batch: iter_time=1.091e-04, loss_ctc=70.602, loss_att=57.368, acc=0.729, loss=61.355, grad_norm=6.538, loss_scale=1.000, learning_rate=6.014e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:30:27,876 (deepspeed_trainer:228) INFO: 45epoch:train:3601-3700batch: iter_time=1.090e-04, loss_ctc=65.786, loss_att=50.299, acc=0.735, loss=54.928, grad_norm=5.431, loss_scale=1.000, learning_rate=6.014e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 02:30:50,211 (multiple_iter_factory:32) INFO: Building 2th iter-factory... [cnode7-018:0/16] 2024-12-11 02:31:17,114 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 02:31:33,198 (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: ) [cnode7-018:0/16] 2024-12-11 02:31:33,198 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.2, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 02:31:33,200 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 02:32:00,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:01,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:01,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:00,408] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,001] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,077] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:01,324] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:01,344] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,594] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,607] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:01,753] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,004] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,048] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,212] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:02,406] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,033] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:49,177] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,384] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,736] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,092] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:50,953] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,856] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,876] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:51,985] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:52,026] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:32:54,490] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:37,275] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:38,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:38,696] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:38,735] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:39,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:39,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:38,818] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:39,175] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:40,363] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:40,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:39,727] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:39,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:40,536] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:41,133] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:45,213] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:33:45,071] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:25,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:26,911] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:27,118] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:26,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:27,711] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:27,825] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:28,914] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:28,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:28,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:28,476] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:29,524] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:28,661] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:29,090] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:29,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:35,101] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:34:35,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 02:35:21,014 (deepspeed_trainer:228) INFO: 45epoch:train:3701-3800batch: iter_time=2.493, loss_ctc=67.246, loss_att=52.106, acc=0.738, loss=56.653, grad_norm=5.581, loss_scale=1.000, learning_rate=6.013e-05, step_time=0.439 [cnode7-018:0/16] 2024-12-11 02:35:58,194 (deepspeed_trainer:228) INFO: 45epoch:train:3801-3900batch: iter_time=1.079e-04, loss_ctc=74.856, loss_att=55.262, acc=0.725, loss=61.121, grad_norm=6.294, loss_scale=1.000, learning_rate=6.013e-05, step_time=0.371 [2024-12-11 02:36:34,978] [INFO] [logging.py:129:log_dist] [Rank 0] step=664000, skipped=0, lr=[np.float64(6.012031593290998e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:36:34,978] [INFO] [timer.py:264:stop] epoch=0/micro_step=64000/global_step=64000, RunningAvgSamplesPerSec=44.143812592053706, CurrSamplesPerSec=47.14716816329219, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:36:34,980 (deepspeed_trainer:228) INFO: 45epoch:train:3901-4000batch: iter_time=1.094e-04, loss_ctc=58.984, loss_att=46.486, acc=0.742, loss=50.256, grad_norm=5.854, loss_scale=1.000, learning_rate=6.012e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:37:11,450 (deepspeed_trainer:228) INFO: 45epoch:train:4001-4100batch: iter_time=1.132e-04, loss_ctc=60.642, loss_att=43.755, acc=0.737, loss=48.823, grad_norm=5.726, loss_scale=1.000, learning_rate=6.012e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 02:37:48,146 (deepspeed_trainer:228) INFO: 45epoch:train:4101-4200batch: iter_time=1.136e-04, loss_ctc=72.835, loss_att=52.222, acc=0.729, loss=58.432, grad_norm=6.945, loss_scale=1.000, learning_rate=6.011e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:38:24,544 (deepspeed_trainer:228) INFO: 45epoch:train:4201-4300batch: iter_time=1.118e-04, loss_ctc=62.613, loss_att=47.358, acc=0.732, loss=51.951, grad_norm=6.462, loss_scale=1.000, learning_rate=6.011e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 02:39:01,724 (deepspeed_trainer:228) INFO: 45epoch:train:4301-4400batch: iter_time=1.107e-04, loss_ctc=68.390, loss_att=54.125, acc=0.723, loss=58.418, grad_norm=6.500, loss_scale=1.000, learning_rate=6.010e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 02:39:39,762 (deepspeed_trainer:228) INFO: 45epoch:train:4401-4500batch: iter_time=1.491e-04, loss_ctc=77.470, loss_att=59.974, acc=0.724, loss=65.225, grad_norm=7.296, loss_scale=1.000, learning_rate=6.010e-05, step_time=0.381 [cnode7-018:0/16] 2024-12-11 02:40:16,820 (deepspeed_trainer:228) INFO: 45epoch:train:4501-4600batch: iter_time=1.087e-04, loss_ctc=60.649, loss_att=42.643, acc=0.741, loss=48.057, grad_norm=5.665, loss_scale=1.000, learning_rate=6.010e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:40:53,708 (deepspeed_trainer:228) INFO: 45epoch:train:4601-4700batch: iter_time=1.098e-04, loss_ctc=65.009, loss_att=55.172, acc=0.720, loss=58.110, grad_norm=6.030, loss_scale=1.000, learning_rate=6.009e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 02:41:30,733 (deepspeed_trainer:228) INFO: 45epoch:train:4701-4800batch: iter_time=1.107e-04, loss_ctc=72.524, loss_att=59.338, acc=0.721, loss=63.292, grad_norm=7.281, loss_scale=1.000, learning_rate=6.009e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:42:08,095 (deepspeed_trainer:228) INFO: 45epoch:train:4801-4900batch: iter_time=1.108e-04, loss_ctc=66.861, loss_att=57.781, acc=0.734, loss=60.473, grad_norm=5.956, loss_scale=1.000, learning_rate=6.008e-05, step_time=0.373 [2024-12-11 02:42:44,883] [INFO] [logging.py:129:log_dist] [Rank 0] step=665000, skipped=0, lr=[np.float64(6.0075095749383116e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:42:44,884] [INFO] [timer.py:264:stop] epoch=0/micro_step=65000/global_step=65000, RunningAvgSamplesPerSec=44.14784935088504, CurrSamplesPerSec=45.464794274230094, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:42:44,885 (deepspeed_trainer:228) INFO: 45epoch:train:4901-5000batch: iter_time=1.098e-04, loss_ctc=60.051, loss_att=47.539, acc=0.730, loss=51.303, grad_norm=6.522, loss_scale=1.000, learning_rate=6.008e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:43:22,022 (deepspeed_trainer:228) INFO: 45epoch:train:5001-5100batch: iter_time=1.118e-04, loss_ctc=68.234, loss_att=50.323, acc=0.730, loss=55.725, grad_norm=6.699, loss_scale=1.000, learning_rate=6.007e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 02:43:58,633 (deepspeed_trainer:228) INFO: 45epoch:train:5101-5200batch: iter_time=1.096e-04, loss_ctc=65.237, loss_att=48.398, acc=0.745, loss=53.459, grad_norm=5.604, loss_scale=1.000, learning_rate=6.007e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:44:35,372 (deepspeed_trainer:228) INFO: 45epoch:train:5201-5300batch: iter_time=1.093e-04, loss_ctc=74.322, loss_att=54.945, acc=0.725, loss=60.743, grad_norm=6.299, loss_scale=1.000, learning_rate=6.006e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:45:12,132 (deepspeed_trainer:228) INFO: 45epoch:train:5301-5400batch: iter_time=1.120e-04, loss_ctc=68.415, loss_att=52.364, acc=0.726, loss=57.191, grad_norm=6.339, loss_scale=1.000, learning_rate=6.006e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 02:45:49,212 (deepspeed_trainer:228) INFO: 45epoch:train:5401-5500batch: iter_time=1.152e-04, loss_ctc=72.548, loss_att=56.453, acc=0.740, loss=61.296, grad_norm=6.012, loss_scale=1.000, learning_rate=6.005e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 02:46:26,074 (deepspeed_trainer:228) INFO: 45epoch:train:5501-5600batch: iter_time=1.158e-04, loss_ctc=58.155, loss_att=45.809, acc=0.734, loss=49.536, grad_norm=5.573, loss_scale=1.000, learning_rate=6.005e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 02:46:38,695 (multiple_iter_factory:32) INFO: Building 3th iter-factory... [cnode7-018:0/16] 2024-12-11 02:47:05,827 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 02:47:22,608 (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: ) [cnode7-018:0/16] 2024-12-11 02:47:22,608 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.1, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 02:47:22,610 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 02:47:50,256] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:50,059] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:50,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:51,049] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:50,492] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:51,444] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:50,610] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:51,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:50,886] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:52,093] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:52,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:51,308] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:51,783] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:52,737] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:52,066] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:47:52,348] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:39,364] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:39,434] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,705] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,107] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,439] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,576] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,454] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:40,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,570] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,608] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,079] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:41,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:44,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:48:45,036] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:27,623] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:28,210] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:29,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:29,712] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:29,875] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,789] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,884] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,056] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:31,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,554] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,721] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:30,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:33,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:35,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:49:35,618] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:16,117] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:17,662] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:16,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:18,739] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:18,386] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:18,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:18,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:19,878] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:19,974] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:19,317] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:20,309] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:19,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:20,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:25,487] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:25,821] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 02:50:25,921] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 02:51:19,442 (deepspeed_trainer:228) INFO: 45epoch:train:5601-5700batch: iter_time=2.479, loss_ctc=73.224, loss_att=55.348, acc=0.733, loss=60.706, grad_norm=6.851, loss_scale=1.000, learning_rate=6.005e-05, step_time=0.455 [cnode7-018:0/16] 2024-12-11 02:51:56,987 (deepspeed_trainer:228) INFO: 45epoch:train:5701-5800batch: iter_time=1.061e-04, loss_ctc=70.373, loss_att=57.529, acc=0.715, loss=61.366, grad_norm=6.031, loss_scale=1.000, learning_rate=6.004e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-11 02:52:33,719 (deepspeed_trainer:228) INFO: 45epoch:train:5801-5900batch: iter_time=1.079e-04, loss_ctc=58.862, loss_att=45.380, acc=0.744, loss=49.449, grad_norm=5.959, loss_scale=1.000, learning_rate=6.004e-05, step_time=0.367 [2024-12-11 02:53:10,244] [INFO] [logging.py:129:log_dist] [Rank 0] step=666000, skipped=0, lr=[np.float64(6.00299774512327e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:53:10,244] [INFO] [timer.py:264:stop] epoch=0/micro_step=66000/global_step=66000, RunningAvgSamplesPerSec=44.1375241246858, CurrSamplesPerSec=46.9678287446918, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:53:10,246 (deepspeed_trainer:228) INFO: 45epoch:train:5901-6000batch: iter_time=1.077e-04, loss_ctc=60.017, loss_att=43.031, acc=0.736, loss=48.134, grad_norm=6.040, loss_scale=1.000, learning_rate=6.003e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 02:53:47,158 (deepspeed_trainer:228) INFO: 45epoch:train:6001-6100batch: iter_time=1.089e-04, loss_ctc=73.014, loss_att=52.523, acc=0.721, loss=58.687, grad_norm=7.145, loss_scale=1.000, learning_rate=6.003e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 02:54:24,399 (deepspeed_trainer:228) INFO: 45epoch:train:6101-6200batch: iter_time=1.123e-04, loss_ctc=64.715, loss_att=46.584, acc=0.730, loss=52.034, grad_norm=6.137, loss_scale=1.000, learning_rate=6.002e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 02:55:01,033 (deepspeed_trainer:228) INFO: 45epoch:train:6201-6300batch: iter_time=1.098e-04, loss_ctc=70.614, loss_att=58.306, acc=0.713, loss=62.016, grad_norm=6.972, loss_scale=1.000, learning_rate=6.002e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 02:55:37,539 (deepspeed_trainer:228) INFO: 45epoch:train:6301-6400batch: iter_time=1.111e-04, loss_ctc=70.132, loss_att=52.093, acc=0.737, loss=57.487, grad_norm=7.108, loss_scale=1.000, learning_rate=6.001e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 02:56:14,519 (deepspeed_trainer:228) INFO: 45epoch:train:6401-6500batch: iter_time=1.087e-04, loss_ctc=64.797, loss_att=49.451, acc=0.729, loss=54.057, grad_norm=6.196, loss_scale=1.000, learning_rate=6.001e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 02:56:51,054 (deepspeed_trainer:228) INFO: 45epoch:train:6501-6600batch: iter_time=1.090e-04, loss_ctc=62.378, loss_att=53.388, acc=0.722, loss=56.083, grad_norm=6.823, loss_scale=1.000, learning_rate=6.001e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 02:57:27,601 (deepspeed_trainer:228) INFO: 45epoch:train:6601-6700batch: iter_time=1.091e-04, loss_ctc=74.760, loss_att=63.122, acc=0.716, loss=66.624, grad_norm=6.806, loss_scale=1.000, learning_rate=6.000e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 02:58:04,080 (deepspeed_trainer:228) INFO: 45epoch:train:6701-6800batch: iter_time=1.092e-04, loss_ctc=61.101, loss_att=48.422, acc=0.730, loss=52.255, grad_norm=6.399, loss_scale=1.000, learning_rate=6.000e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 02:58:40,482 (deepspeed_trainer:228) INFO: 45epoch:train:6801-6900batch: iter_time=1.107e-04, loss_ctc=63.702, loss_att=50.431, acc=0.726, loss=54.401, grad_norm=6.532, loss_scale=1.000, learning_rate=5.999e-05, step_time=0.364 [2024-12-11 02:59:16,875] [INFO] [logging.py:129:log_dist] [Rank 0] step=667000, skipped=0, lr=[np.float64(5.998496065643675e-05)], mom=[[0.9, 0.98]] [2024-12-11 02:59:16,876] [INFO] [timer.py:264:stop] epoch=0/micro_step=67000/global_step=67000, RunningAvgSamplesPerSec=44.14724239189936, CurrSamplesPerSec=46.71539108919698, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 02:59:16,877 (deepspeed_trainer:228) INFO: 45epoch:train:6901-7000batch: iter_time=1.092e-04, loss_ctc=64.664, loss_att=46.320, acc=0.733, loss=51.824, grad_norm=5.583, loss_scale=1.000, learning_rate=5.999e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 02:59:53,466 (deepspeed_trainer:228) INFO: 45epoch:train:7001-7100batch: iter_time=1.105e-04, loss_ctc=69.131, loss_att=52.838, acc=0.735, loss=57.714, grad_norm=5.363, loss_scale=1.000, learning_rate=5.998e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 03:00:29,787 (deepspeed_trainer:228) INFO: 45epoch:train:7101-7200batch: iter_time=1.111e-04, loss_ctc=71.469, loss_att=51.014, acc=0.724, loss=57.160, grad_norm=6.950, loss_scale=1.000, learning_rate=5.998e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 03:01:06,487 (deepspeed_trainer:228) INFO: 45epoch:train:7201-7300batch: iter_time=1.101e-04, loss_ctc=67.507, loss_att=53.023, acc=0.730, loss=57.351, grad_norm=5.834, loss_scale=1.000, learning_rate=5.997e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:01:43,360 (deepspeed_trainer:228) INFO: 45epoch:train:7301-7400batch: iter_time=1.095e-04, loss_ctc=71.546, loss_att=58.824, acc=0.728, loss=62.618, grad_norm=6.240, loss_scale=1.000, learning_rate=5.997e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:02:19,512 (deepspeed_trainer:228) INFO: 45epoch:train:7401-7500batch: iter_time=1.075e-04, loss_ctc=62.391, loss_att=45.391, acc=0.734, loss=50.486, grad_norm=5.723, loss_scale=1.000, learning_rate=5.996e-05, step_time=0.361 [cnode7-018:0/16] 2024-12-11 03:02:22,973 (multiple_iter_factory:32) INFO: Building 4th iter-factory... [cnode7-018:0/16] 2024-12-11 03:02:50,023 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 03:03:06,505 (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: ) [cnode7-018:0/16] 2024-12-11 03:03:06,505 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.7, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 03:03:06,508 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 03:03:32,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:32,940] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:33,429] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:33,740] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,507] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,710] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,617] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,935] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,848] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:34,904] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:35,376] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:35,717] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:03:35,769] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:21,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:21,832] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:21,916] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:22,086] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:22,947] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:22,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:23,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:23,140] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:23,533] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:23,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:23,979] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:24,104] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:24,621] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:24,879] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:24,977] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:04:25,244] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:09,633] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:10,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:10,658] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:11,326] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:11,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:11,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,030] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,438] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,352] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,496] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,801] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,891] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:12,787] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:13,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:14,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:16,934] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:56,630] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:58,414] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:58,632] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:58,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:05:59,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:00,341] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:00,861] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:00,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:00,777] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:00,999] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:01,084] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:01,310] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:01,436] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:02,181] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:02,829] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:06:07,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 03:07:11,051 (deepspeed_trainer:228) INFO: 45epoch:train:7501-7600batch: iter_time=2.476, loss_ctc=71.184, loss_att=53.727, acc=0.728, loss=58.973, grad_norm=6.337, loss_scale=1.000, learning_rate=5.996e-05, step_time=0.439 [cnode7-018:0/16] 2024-12-11 03:07:48,261 (deepspeed_trainer:228) INFO: 45epoch:train:7601-7700batch: iter_time=1.091e-04, loss_ctc=65.987, loss_att=49.796, acc=0.735, loss=54.641, grad_norm=6.166, loss_scale=1.000, learning_rate=5.996e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 03:08:25,145 (deepspeed_trainer:228) INFO: 45epoch:train:7701-7800batch: iter_time=1.088e-04, loss_ctc=59.271, loss_att=44.123, acc=0.742, loss=48.657, grad_norm=5.263, loss_scale=1.000, learning_rate=5.995e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:09:01,699 (deepspeed_trainer:228) INFO: 45epoch:train:7801-7900batch: iter_time=1.071e-04, loss_ctc=63.232, loss_att=44.908, acc=0.735, loss=50.383, grad_norm=5.706, loss_scale=1.000, learning_rate=5.995e-05, step_time=0.365 [2024-12-11 03:09:38,337] [INFO] [logging.py:129:log_dist] [Rank 0] step=668000, skipped=0, lr=[np.float64(5.99400449849757e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:09:38,337] [INFO] [timer.py:264:stop] epoch=0/micro_step=68000/global_step=68000, RunningAvgSamplesPerSec=44.143481878612334, CurrSamplesPerSec=46.74389530573103, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:09:38,339 (deepspeed_trainer:228) INFO: 45epoch:train:7901-8000batch: iter_time=1.058e-04, loss_ctc=70.312, loss_att=52.975, acc=0.715, loss=58.196, grad_norm=6.964, loss_scale=1.000, learning_rate=5.994e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:10:14,712 (deepspeed_trainer:228) INFO: 45epoch:train:8001-8100batch: iter_time=1.088e-04, loss_ctc=62.438, loss_att=42.907, acc=0.732, loss=48.759, grad_norm=5.816, loss_scale=1.000, learning_rate=5.994e-05, step_time=0.363 [cnode7-018:0/16] 2024-12-11 03:10:52,026 (deepspeed_trainer:228) INFO: 45epoch:train:8101-8200batch: iter_time=1.124e-04, loss_ctc=77.213, loss_att=63.691, acc=0.715, loss=67.761, grad_norm=6.731, loss_scale=1.000, learning_rate=5.993e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 03:11:28,772 (deepspeed_trainer:228) INFO: 45epoch:train:8201-8300batch: iter_time=1.109e-04, loss_ctc=65.639, loss_att=46.129, acc=0.743, loss=51.979, grad_norm=6.205, loss_scale=1.000, learning_rate=5.993e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:12:05,358 (deepspeed_trainer:228) INFO: 45epoch:train:8301-8400batch: iter_time=1.101e-04, loss_ctc=64.057, loss_att=52.277, acc=0.720, loss=55.814, grad_norm=6.014, loss_scale=1.000, learning_rate=5.992e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:12:42,060 (deepspeed_trainer:228) INFO: 45epoch:train:8401-8500batch: iter_time=1.098e-04, loss_ctc=65.047, loss_att=53.915, acc=0.729, loss=57.232, grad_norm=6.686, loss_scale=1.000, learning_rate=5.992e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:13:18,816 (deepspeed_trainer:228) INFO: 45epoch:train:8501-8600batch: iter_time=1.054e-04, loss_ctc=74.362, loss_att=62.817, acc=0.715, loss=66.263, grad_norm=6.588, loss_scale=1.000, learning_rate=5.992e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:13:55,454 (deepspeed_trainer:228) INFO: 45epoch:train:8601-8700batch: iter_time=1.073e-04, loss_ctc=57.526, loss_att=45.197, acc=0.727, loss=48.882, grad_norm=6.021, loss_scale=1.000, learning_rate=5.991e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:14:32,054 (deepspeed_trainer:228) INFO: 45epoch:train:8701-8800batch: iter_time=1.059e-04, loss_ctc=67.118, loss_att=52.583, acc=0.729, loss=56.934, grad_norm=5.938, loss_scale=1.000, learning_rate=5.991e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 03:15:08,406 (deepspeed_trainer:228) INFO: 45epoch:train:8801-8900batch: iter_time=1.089e-04, loss_ctc=65.696, loss_att=46.992, acc=0.731, loss=52.572, grad_norm=5.609, loss_scale=1.000, learning_rate=5.990e-05, step_time=0.363 [2024-12-11 03:15:45,051] [INFO] [logging.py:129:log_dist] [Rank 0] step=669000, skipped=0, lr=[np.float64(5.9895230058818785e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:15:45,051] [INFO] [timer.py:264:stop] epoch=0/micro_step=69000/global_step=69000, RunningAvgSamplesPerSec=44.15239892277836, CurrSamplesPerSec=45.166846406479664, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:15:45,053 (deepspeed_trainer:228) INFO: 45epoch:train:8901-9000batch: iter_time=1.080e-04, loss_ctc=70.881, loss_att=53.889, acc=0.738, loss=58.970, grad_norm=6.280, loss_scale=1.000, learning_rate=5.990e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:16:21,529 (deepspeed_trainer:228) INFO: 45epoch:train:9001-9100batch: iter_time=1.059e-04, loss_ctc=70.266, loss_att=49.494, acc=0.720, loss=55.697, grad_norm=6.324, loss_scale=1.000, learning_rate=5.989e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 03:16:57,940 (deepspeed_trainer:228) INFO: 45epoch:train:9101-9200batch: iter_time=1.088e-04, loss_ctc=67.452, loss_att=51.367, acc=0.737, loss=56.172, grad_norm=5.968, loss_scale=1.000, learning_rate=5.989e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 03:17:34,859 (deepspeed_trainer:228) INFO: 45epoch:train:9201-9300batch: iter_time=1.073e-04, loss_ctc=67.830, loss_att=56.460, acc=0.728, loss=59.853, grad_norm=5.783, loss_scale=1.000, learning_rate=5.988e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:18:05,154 (multiple_iter_factory:32) INFO: Building 5th iter-factory... [cnode7-018:0/16] 2024-12-11 03:18:32,260 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 03:18:47,776 (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: ) [cnode7-018:0/16] 2024-12-11 03:18:47,777 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.4, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 03:18:47,779 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 03:19:15,074] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:15,514] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:15,765] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,043] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,447] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,931] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:16,841] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:17,130] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:17,300] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:17,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:17,351] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:17,756] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:18,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:19:18,680] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:04,283] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:04,668] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,034] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,320] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,469] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,505] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,571] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:05,836] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:06,240] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:06,337] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:06,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:06,793] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:07,098] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:07,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:07,378] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:07,732] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:52,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:52,760] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:53,759] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:54,304] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:54,474] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,185] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,471] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,538] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,551] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:55,708] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:56,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:56,758] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:57,062] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:20:58,593] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:00,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:39,725] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:40,562] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:42,890] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:43,047] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:43,702] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:44,088] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:44,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:44,298] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:44,988] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:45,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:45,527] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:45,854] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:46,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:46,079] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:49,147] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:21:52,431] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 03:22:27,207 (deepspeed_trainer:228) INFO: 45epoch:train:9301-9400batch: iter_time=2.493, loss_ctc=66.440, loss_att=51.035, acc=0.734, loss=55.669, grad_norm=5.816, loss_scale=1.000, learning_rate=5.988e-05, step_time=0.430 [cnode7-018:0/16] 2024-12-11 03:23:04,140 (deepspeed_trainer:228) INFO: 45epoch:train:9401-9500batch: iter_time=1.083e-04, loss_ctc=69.597, loss_att=52.718, acc=0.727, loss=57.757, grad_norm=6.170, loss_scale=1.000, learning_rate=5.988e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:23:40,829 (deepspeed_trainer:228) INFO: 45epoch:train:9501-9600batch: iter_time=1.073e-04, loss_ctc=64.038, loss_att=49.922, acc=0.743, loss=54.153, grad_norm=5.340, loss_scale=1.000, learning_rate=5.987e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:24:17,641 (deepspeed_trainer:228) INFO: 45epoch:train:9601-9700batch: iter_time=1.111e-04, loss_ctc=60.525, loss_att=44.662, acc=0.743, loss=49.409, grad_norm=5.742, loss_scale=1.000, learning_rate=5.987e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:24:54,468 (deepspeed_trainer:228) INFO: 45epoch:train:9701-9800batch: iter_time=1.130e-04, loss_ctc=68.511, loss_att=48.722, acc=0.733, loss=54.650, grad_norm=6.147, loss_scale=1.000, learning_rate=5.986e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:25:30,935 (deepspeed_trainer:228) INFO: 45epoch:train:9801-9900batch: iter_time=1.150e-04, loss_ctc=62.066, loss_att=48.013, acc=0.729, loss=52.209, grad_norm=5.970, loss_scale=1.000, learning_rate=5.986e-05, step_time=0.365 [2024-12-11 03:26:07,815] [INFO] [logging.py:129:log_dist] [Rank 0] step=670000, skipped=0, lr=[np.float64(5.9850515501910837e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:26:07,816] [INFO] [timer.py:264:stop] epoch=0/micro_step=70000/global_step=70000, RunningAvgSamplesPerSec=44.14935980899941, CurrSamplesPerSec=43.93700583143512, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:26:07,817 (deepspeed_trainer:228) INFO: 45epoch:train:9901-10000batch: iter_time=1.094e-04, loss_ctc=66.533, loss_att=51.156, acc=0.720, loss=55.786, grad_norm=5.893, loss_scale=1.000, learning_rate=5.985e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:26:45,258 (deepspeed_trainer:228) INFO: 45epoch:train:10001-10100batch: iter_time=1.095e-04, loss_ctc=77.397, loss_att=60.513, acc=0.734, loss=65.589, grad_norm=6.836, loss_scale=1.000, learning_rate=5.985e-05, step_time=0.374 [cnode7-018:0/16] 2024-12-11 03:27:22,217 (deepspeed_trainer:228) INFO: 45epoch:train:10101-10200batch: iter_time=1.097e-04, loss_ctc=62.552, loss_att=45.122, acc=0.741, loss=50.345, grad_norm=6.110, loss_scale=1.000, learning_rate=5.984e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:27:59,074 (deepspeed_trainer:228) INFO: 45epoch:train:10201-10300batch: iter_time=1.097e-04, loss_ctc=64.376, loss_att=56.604, acc=0.720, loss=58.909, grad_norm=6.108, loss_scale=1.000, learning_rate=5.984e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:28:35,905 (deepspeed_trainer:228) INFO: 45epoch:train:10301-10400batch: iter_time=1.102e-04, loss_ctc=71.429, loss_att=56.329, acc=0.728, loss=60.897, grad_norm=6.765, loss_scale=1.000, learning_rate=5.983e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:29:12,971 (deepspeed_trainer:228) INFO: 45epoch:train:10401-10500batch: iter_time=1.098e-04, loss_ctc=66.614, loss_att=57.996, acc=0.731, loss=60.581, grad_norm=5.862, loss_scale=1.000, learning_rate=5.983e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 03:29:50,326 (deepspeed_trainer:228) INFO: 45epoch:train:10501-10600batch: iter_time=1.102e-04, loss_ctc=59.696, loss_att=49.569, acc=0.731, loss=52.608, grad_norm=6.245, loss_scale=1.000, learning_rate=5.983e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 03:30:27,282 (deepspeed_trainer:228) INFO: 45epoch:train:10601-10700batch: iter_time=1.114e-04, loss_ctc=64.080, loss_att=47.208, acc=0.738, loss=52.252, grad_norm=6.108, loss_scale=1.000, learning_rate=5.982e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 03:31:04,110 (deepspeed_trainer:228) INFO: 45epoch:train:10701-10800batch: iter_time=1.099e-04, loss_ctc=68.038, loss_att=50.013, acc=0.740, loss=55.422, grad_norm=5.670, loss_scale=1.000, learning_rate=5.982e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:31:40,717 (deepspeed_trainer:228) INFO: 45epoch:train:10801-10900batch: iter_time=1.110e-04, loss_ctc=69.455, loss_att=52.762, acc=0.728, loss=57.770, grad_norm=6.429, loss_scale=1.000, learning_rate=5.981e-05, step_time=0.365 [2024-12-11 03:32:17,260] [INFO] [logging.py:129:log_dist] [Rank 0] step=671000, skipped=0, lr=[np.float64(5.980590094015893e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:32:17,260] [INFO] [timer.py:264:stop] epoch=0/micro_step=71000/global_step=71000, RunningAvgSamplesPerSec=44.153784358343415, CurrSamplesPerSec=45.47477612155349, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:32:17,262 (deepspeed_trainer:228) INFO: 45epoch:train:10901-11000batch: iter_time=1.104e-04, loss_ctc=67.507, loss_att=48.794, acc=0.737, loss=54.375, grad_norm=6.400, loss_scale=1.000, learning_rate=5.981e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 03:32:54,105 (deepspeed_trainer:228) INFO: 45epoch:train:11001-11100batch: iter_time=1.104e-04, loss_ctc=69.729, loss_att=57.020, acc=0.732, loss=60.854, grad_norm=6.100, loss_scale=1.000, learning_rate=5.980e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:33:30,785 (deepspeed_trainer:228) INFO: 45epoch:train:11101-11200batch: iter_time=1.124e-04, loss_ctc=65.904, loss_att=50.245, acc=0.737, loss=54.945, grad_norm=5.486, loss_scale=1.000, learning_rate=5.980e-05, step_time=0.367 [cnode7-018:0/16] 2024-12-11 03:33:52,119 (multiple_iter_factory:32) INFO: Building 6th iter-factory... [cnode7-018:0/16] 2024-12-11 03:34:19,692 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 03:34:35,997 (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: ) [cnode7-018:0/16] 2024-12-11 03:34:35,997 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.6, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 03:34:35,999 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 03:35:02,196] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:02,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:03,200] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:03,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:03,555] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:03,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,009] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,142] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,161] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,508] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,645] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,790] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,722] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:04,799] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:05,038] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:05,358] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:51,284] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:51,286] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:51,592] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:51,887] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:51,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:52,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:52,529] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:52,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:52,966] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:53,017] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:53,540] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:53,850] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:53,990] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:54,226] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:56,870] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:35:57,638] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:39,603] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:38,099] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:38,138] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:40,095] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:39,163] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:39,243] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:40,892] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:39,738] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:41,734] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:40,501] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:42,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:42,641] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:42,920] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:46,268] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:46,122] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:36:46,928] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:27,559] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:26,761] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:28,498] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:27,116] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:27,488] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:29,402] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:27,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:28,132] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:28,405] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:30,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:30,466] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:30,549] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:31,453] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:36,260] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:36,313] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:37:37,410] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 03:38:26,141 (deepspeed_trainer:228) INFO: 45epoch:train:11201-11300batch: iter_time=2.484, loss_ctc=66.078, loss_att=51.572, acc=0.740, loss=55.931, grad_norm=5.719, loss_scale=1.000, learning_rate=5.979e-05, step_time=0.469 [cnode7-018:0/16] 2024-12-11 03:39:03,616 (deepspeed_trainer:228) INFO: 45epoch:train:11301-11400batch: iter_time=1.042e-04, loss_ctc=73.037, loss_att=54.185, acc=0.731, loss=59.826, grad_norm=6.298, loss_scale=1.000, learning_rate=5.979e-05, step_time=0.375 [cnode7-018:0/16] 2024-12-11 03:39:40,829 (deepspeed_trainer:228) INFO: 45epoch:train:11401-11500batch: iter_time=1.052e-04, loss_ctc=58.565, loss_att=46.113, acc=0.745, loss=49.858, grad_norm=6.077, loss_scale=1.000, learning_rate=5.979e-05, step_time=0.372 [cnode7-018:0/16] 2024-12-11 03:40:17,397 (deepspeed_trainer:228) INFO: 45epoch:train:11501-11600batch: iter_time=1.085e-04, loss_ctc=60.474, loss_att=43.250, acc=0.740, loss=48.420, grad_norm=6.153, loss_scale=1.000, learning_rate=5.978e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 03:40:54,212 (deepspeed_trainer:228) INFO: 45epoch:train:11601-11700batch: iter_time=1.068e-04, loss_ctc=71.824, loss_att=51.914, acc=0.729, loss=57.858, grad_norm=6.824, loss_scale=1.000, learning_rate=5.978e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:41:31,174 (deepspeed_trainer:228) INFO: 45epoch:train:11701-11800batch: iter_time=1.128e-04, loss_ctc=61.803, loss_att=47.127, acc=0.733, loss=51.555, grad_norm=6.225, loss_scale=1.000, learning_rate=5.977e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:42:08,389 (deepspeed_trainer:228) INFO: 45epoch:train:11801-11900batch: iter_time=1.092e-04, loss_ctc=67.545, loss_att=54.080, acc=0.725, loss=58.089, grad_norm=6.784, loss_scale=1.000, learning_rate=5.977e-05, step_time=0.371 [2024-12-11 03:42:45,611] [INFO] [logging.py:129:log_dist] [Rank 0] step=672000, skipped=0, lr=[np.float64(5.976138600141926e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:42:45,611] [INFO] [timer.py:264:stop] epoch=0/micro_step=72000/global_step=72000, RunningAvgSamplesPerSec=44.13984261445798, CurrSamplesPerSec=44.870948043754645, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:42:45,613 (deepspeed_trainer:228) INFO: 45epoch:train:11901-12000batch: iter_time=1.086e-04, loss_ctc=76.662, loss_att=59.516, acc=0.727, loss=64.662, grad_norm=7.244, loss_scale=1.000, learning_rate=5.976e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 03:43:22,000 (deepspeed_trainer:228) INFO: 45epoch:train:12001-12100batch: iter_time=1.092e-04, loss_ctc=59.829, loss_att=42.285, acc=0.745, loss=47.553, grad_norm=5.665, loss_scale=1.000, learning_rate=5.976e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 03:43:58,656 (deepspeed_trainer:228) INFO: 45epoch:train:12101-12200batch: iter_time=1.068e-04, loss_ctc=63.839, loss_att=54.705, acc=0.724, loss=57.447, grad_norm=6.528, loss_scale=1.000, learning_rate=5.975e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:44:35,596 (deepspeed_trainer:228) INFO: 45epoch:train:12201-12300batch: iter_time=1.082e-04, loss_ctc=71.990, loss_att=58.945, acc=0.726, loss=62.871, grad_norm=7.020, loss_scale=1.000, learning_rate=5.975e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:45:12,477 (deepspeed_trainer:228) INFO: 45epoch:train:12301-12400batch: iter_time=1.064e-04, loss_ctc=66.463, loss_att=57.077, acc=0.735, loss=59.872, grad_norm=5.697, loss_scale=1.000, learning_rate=5.975e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:45:48,989 (deepspeed_trainer:228) INFO: 45epoch:train:12401-12500batch: iter_time=1.075e-04, loss_ctc=59.584, loss_att=47.002, acc=0.732, loss=50.782, grad_norm=6.128, loss_scale=1.000, learning_rate=5.974e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 03:46:25,959 (deepspeed_trainer:228) INFO: 45epoch:train:12501-12600batch: iter_time=1.075e-04, loss_ctc=66.771, loss_att=49.591, acc=0.733, loss=54.745, grad_norm=6.303, loss_scale=1.000, learning_rate=5.974e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:47:02,336 (deepspeed_trainer:228) INFO: 45epoch:train:12601-12700batch: iter_time=1.066e-04, loss_ctc=64.249, loss_att=47.905, acc=0.746, loss=52.828, grad_norm=5.350, loss_scale=1.000, learning_rate=5.973e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 03:47:38,783 (deepspeed_trainer:228) INFO: 45epoch:train:12701-12800batch: iter_time=1.082e-04, loss_ctc=72.618, loss_att=54.726, acc=0.726, loss=60.110, grad_norm=6.643, loss_scale=1.000, learning_rate=5.973e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 03:48:15,160 (deepspeed_trainer:228) INFO: 45epoch:train:12801-12900batch: iter_time=1.081e-04, loss_ctc=67.608, loss_att=52.042, acc=0.729, loss=56.727, grad_norm=6.405, loss_scale=1.000, learning_rate=5.972e-05, step_time=0.363 [2024-12-11 03:48:52,536] [INFO] [logging.py:129:log_dist] [Rank 0] step=673000, skipped=0, lr=[np.float64(5.971697031548412e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:48:52,537] [INFO] [timer.py:264:stop] epoch=0/micro_step=73000/global_step=73000, RunningAvgSamplesPerSec=44.147856677259604, CurrSamplesPerSec=43.504238703000816, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:48:52,538 (deepspeed_trainer:228) INFO: 45epoch:train:12901-13000batch: iter_time=1.077e-04, loss_ctc=71.104, loss_att=55.808, acc=0.742, loss=60.397, grad_norm=5.973, loss_scale=1.000, learning_rate=5.972e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 03:49:28,810 (deepspeed_trainer:228) INFO: 45epoch:train:13001-13100batch: iter_time=1.079e-04, loss_ctc=57.626, loss_att=45.548, acc=0.736, loss=49.185, grad_norm=5.437, loss_scale=1.000, learning_rate=5.971e-05, step_time=0.362 [cnode7-018:0/16] 2024-12-11 03:49:41,626 (multiple_iter_factory:32) INFO: Building 7th iter-factory... [cnode7-018:0/16] 2024-12-11 03:50:09,152 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4') [cnode7-018:0/16] 2024-12-11 03:50:24,682 (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: ) [cnode7-018:0/16] 2024-12-11 03:50:24,682 (abs_task:1808) INFO: [train] Batch sampler: SortedBatchSampler(N-batch=28521, batch_size=256, shape_file=exp_owsm/s2t_stats_raw_bpe50000/splits8/speech_shape/split.3, sort_in_batch=descending, sort_batch=descending) [cnode7-018:0/16] 2024-12-11 03:50:24,684 (abs_task:1809) INFO: [train] mini-batch sizes summary: N-batch=28521, mean=256.0, min=256, max=257 [2024-12-11 03:50:52,851] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:52,893] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:53,176] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:53,987] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:52,463] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:52,803] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:52,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:54,606] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:54,897] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:55,278] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:53,703] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:53,775] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:55,416] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:53,898] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:54,091] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:50:54,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:41,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:42,215] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:42,265] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:42,619] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:41,640] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:41,946] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:42,400] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:43,907] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:42,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:43,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:44,069] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:43,051] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:44,448] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:43,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:43,674] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:51:46,568] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:30,742] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:31,021] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,239] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,293] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,745] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:31,581] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,925] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,190] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,202] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:33,472] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:33,518] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:32,631] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:33,836] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:34,225] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:33,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:52:38,952] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:20,049] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:21,428] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:20,748] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:21,930] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:22,058] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:20,918] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:21,238] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:22,983] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:21,868] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:23,040] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:23,155] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:22,050] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:23,451] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:22,374] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:23,547] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 03:53:29,139] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [cnode7-018:0/16] 2024-12-11 03:54:23,368 (deepspeed_trainer:228) INFO: 45epoch:train:13101-13200batch: iter_time=2.529, loss_ctc=72.332, loss_att=55.075, acc=0.733, loss=60.242, grad_norm=6.398, loss_scale=1.000, learning_rate=5.971e-05, step_time=0.421 [cnode7-018:0/16] 2024-12-11 03:55:00,529 (deepspeed_trainer:228) INFO: 45epoch:train:13201-13300batch: iter_time=1.074e-04, loss_ctc=69.575, loss_att=57.007, acc=0.717, loss=60.812, grad_norm=6.121, loss_scale=1.000, learning_rate=5.971e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 03:55:37,573 (deepspeed_trainer:228) INFO: 45epoch:train:13301-13400batch: iter_time=1.137e-04, loss_ctc=58.997, loss_att=45.403, acc=0.745, loss=49.467, grad_norm=5.463, loss_scale=1.000, learning_rate=5.970e-05, step_time=0.371 [cnode7-018:0/16] 2024-12-11 03:56:14,458 (deepspeed_trainer:228) INFO: 45epoch:train:13401-13500batch: iter_time=1.130e-04, loss_ctc=59.823, loss_att=43.318, acc=0.734, loss=48.247, grad_norm=5.724, loss_scale=1.000, learning_rate=5.970e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 03:56:51,297 (deepspeed_trainer:228) INFO: 45epoch:train:13501-13600batch: iter_time=1.120e-04, loss_ctc=71.938, loss_att=51.649, acc=0.723, loss=57.737, grad_norm=6.716, loss_scale=1.000, learning_rate=5.969e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 03:57:28,354 (deepspeed_trainer:228) INFO: 45epoch:train:13601-13700batch: iter_time=1.139e-04, loss_ctc=63.903, loss_att=46.705, acc=0.729, loss=51.836, grad_norm=5.725, loss_scale=1.000, learning_rate=5.969e-05, step_time=0.370 [cnode7-018:0/16] 2024-12-11 03:58:05,674 (deepspeed_trainer:228) INFO: 45epoch:train:13701-13800batch: iter_time=1.145e-04, loss_ctc=69.612, loss_att=56.648, acc=0.718, loss=60.523, grad_norm=6.714, loss_scale=1.000, learning_rate=5.968e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 03:58:43,119 (deepspeed_trainer:228) INFO: 45epoch:train:13801-13900batch: iter_time=1.175e-04, loss_ctc=70.179, loss_att=52.594, acc=0.737, loss=57.858, grad_norm=6.311, loss_scale=1.000, learning_rate=5.968e-05, step_time=0.374 [2024-12-11 03:59:19,749] [INFO] [logging.py:129:log_dist] [Rank 0] step=674000, skipped=0, lr=[np.float64(5.967265351406893e-05)], mom=[[0.9, 0.98]] [2024-12-11 03:59:19,750] [INFO] [timer.py:264:stop] epoch=0/micro_step=74000/global_step=74000, RunningAvgSamplesPerSec=44.14261293328033, CurrSamplesPerSec=42.04509937277054, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 03:59:19,751 (deepspeed_trainer:228) INFO: 45epoch:train:13901-14000batch: iter_time=1.143e-04, loss_ctc=64.411, loss_att=48.909, acc=0.732, loss=53.564, grad_norm=5.802, loss_scale=1.000, learning_rate=5.967e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 03:59:56,693 (deepspeed_trainer:228) INFO: 45epoch:train:14001-14100batch: iter_time=1.124e-04, loss_ctc=61.424, loss_att=54.093, acc=0.722, loss=56.290, grad_norm=6.516, loss_scale=1.000, learning_rate=5.967e-05, step_time=0.369 [cnode7-018:0/16] 2024-12-11 04:00:33,971 (deepspeed_trainer:228) INFO: 45epoch:train:14101-14200batch: iter_time=1.134e-04, loss_ctc=73.748, loss_att=63.246, acc=0.716, loss=66.381, grad_norm=6.432, loss_scale=1.000, learning_rate=5.967e-05, step_time=0.373 [cnode7-018:0/16] 2024-12-11 04:01:10,841 (deepspeed_trainer:228) INFO: 45epoch:train:14201-14300batch: iter_time=1.117e-04, loss_ctc=60.754, loss_att=48.097, acc=0.731, loss=51.905, grad_norm=5.989, loss_scale=1.000, learning_rate=5.966e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 04:01:47,681 (deepspeed_trainer:228) INFO: 45epoch:train:14301-14400batch: iter_time=1.125e-04, loss_ctc=62.997, loss_att=50.063, acc=0.727, loss=53.931, grad_norm=6.048, loss_scale=1.000, learning_rate=5.966e-05, step_time=0.368 [cnode7-018:0/16] 2024-12-11 04:02:24,315 (deepspeed_trainer:228) INFO: 45epoch:train:14401-14500batch: iter_time=1.138e-04, loss_ctc=65.052, loss_att=46.673, acc=0.733, loss=52.200, grad_norm=6.232, loss_scale=1.000, learning_rate=5.965e-05, step_time=0.366 [cnode7-018:0/16] 2024-12-11 04:03:00,820 (deepspeed_trainer:228) INFO: 45epoch:train:14501-14600batch: iter_time=1.200e-04, loss_ctc=68.937, loss_att=52.438, acc=0.738, loss=57.404, grad_norm=5.547, loss_scale=1.000, learning_rate=5.965e-05, step_time=0.364 [cnode7-018:0/16] 2024-12-11 04:03:37,286 (deepspeed_trainer:228) INFO: 45epoch:train:14601-14700batch: iter_time=1.144e-04, loss_ctc=70.959, loss_att=50.783, acc=0.728, loss=56.821, grad_norm=7.168, loss_scale=1.000, learning_rate=5.964e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 04:04:13,773 (deepspeed_trainer:228) INFO: 45epoch:train:14701-14800batch: iter_time=1.146e-04, loss_ctc=66.847, loss_att=52.850, acc=0.731, loss=57.031, grad_norm=5.882, loss_scale=1.000, learning_rate=5.964e-05, step_time=0.365 [cnode7-018:0/16] 2024-12-11 04:04:50,540 (deepspeed_trainer:228) INFO: 45epoch:train:14801-14900batch: iter_time=1.142e-04, loss_ctc=70.511, loss_att=58.059, acc=0.730, loss=61.823, grad_norm=6.314, loss_scale=1.000, learning_rate=5.964e-05, step_time=0.367 [2024-12-11 04:05:26,747] [INFO] [logging.py:129:log_dist] [Rank 0] step=675000, skipped=0, lr=[np.float64(5.962843523079948e-05)], mom=[[0.9, 0.98]] [2024-12-11 04:05:26,748] [INFO] [timer.py:264:stop] epoch=0/micro_step=75000/global_step=75000, RunningAvgSamplesPerSec=44.150134359824726, CurrSamplesPerSec=44.05214103259612, MemAllocated=2.04GB, MaxMemAllocated=15.97GB [cnode7-018:0/16] 2024-12-11 04:05:26,749 (deepspeed_trainer:228) INFO: 45epoch:train:14901-15000batch: iter_time=1.132e-04, loss_ctc=62.296, loss_att=45.435, acc=0.735, loss=50.481, grad_norm=5.757, loss_scale=1.000, learning_rate=5.963e-05, step_time=0.362 [2024-12-11 04:05:41,223] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,460] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,626] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,682] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,812] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,869] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,919] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:41,772] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,976] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:41,840] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:41,858] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:43,055] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:41,956] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:41,969] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,041] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:42,065] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:57,252] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:58,958] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:57,889] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:58,441] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:58,483] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,251] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:59,083] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:59,134] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,420] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:59,394] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,636] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,659] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:05:59,513] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,767] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:00,828] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:14,810] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:13,845] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:14,187] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:14,713] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:15,933] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:14,980] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:16,188] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:16,491] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:15,346] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:15,678] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:15,863] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:17,136] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:17,421] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:16,295] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:17,964] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:18,010] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:30,427] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:30,169] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:30,470] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:30,820] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:31,294] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:32,644] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:31,948] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:31,961] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:33,160] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:33,203] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:32,564] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:33,936] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:32,794] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:33,963] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:34,331] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:34,372] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) [2024-12-11 04:06:47,065] [INFO] [logging.py:129:log_dist] [Rank 0] [Torch] Checkpoint 45 is about to be saved! [2024-12-11 04:06:47,095] [INFO] [logging.py:129:log_dist] [Rank 0] Saving model checkpoint: exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/mp_rank_00_model_states.pt [2024-12-11 04:06:47,095] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/mp_rank_00_model_states.pt... [2024-12-11 04:06:48,995] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/mp_rank_00_model_states.pt. [2024-12-11 04:06:47,984] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,986] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,157] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,157] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,989] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,158] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,158] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,158] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,989] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,989] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,990] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,159] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,990] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt... [2024-12-11 04:06:47,991] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,161] [INFO] [torch_checkpoint_engine.py:21:save] [Torch] Saving exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt... [2024-12-11 04:06:49,794] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,794] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,794] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:49,927] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,927] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,927] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:49,940] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,940] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,940] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,785] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,785] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,785] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:49,960] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,960] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,961] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:49,964] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,964] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,965] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,798] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,798] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,798] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,799] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,799] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,799] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,804] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,804] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,804] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,810] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,810] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,810] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,814] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,815] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,815] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:49,991] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt. [2024-12-11 04:06:49,996] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt [2024-12-11 04:06:49,996] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:50,007] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt. [2024-12-11 04:06:50,008] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt [2024-12-11 04:06:50,008] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,859] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,859] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,859] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:48,859] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt. [2024-12-11 04:06:48,859] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt [2024-12-11 04:06:48,859] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [2024-12-11 04:06:50,035] [INFO] [torch_checkpoint_engine.py:23:save] [Torch] Saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt. [2024-12-11 04:06:50,035] [INFO] [engine.py:3536:_save_zero_checkpoint] zero checkpoint saved exp_owsm/s2t_train_05b_ds_raw_bpe50000/checkpoint_45/45/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt [2024-12-11 04:06:50,035] [INFO] [torch_checkpoint_engine.py:33:commit] [Torch] Checkpoint 45 is ready now! [cnode7-018:0/16] 2024-12-11 04:06:50,120 (deepspeed_trainer:158) INFO: 45epoch results: [train] iter_time=0.132, loss_ctc=67.020, loss_att=51.621, acc=0.730, loss=56.240, grad_norm=6.223, loss_scale=1.000, learning_rate=5.996e-05, step_time=0.372, time=2 hours, 6 minutes and 17.95 seconds, total_count=675045, gpu_max_cached_mem_GB=28.656, [valid] loss_ctc=3.391, cer_ctc=0.088, loss_att=7.219, acc=0.805, cer=0.412, wer=1.000, loss=6.062, time=1 minute and 11.95 seconds, total_count=45, gpu_max_cached_mem_GB=28.656 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 2 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 5 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 1 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 7 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 6 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 4 cnode7-018:1407096:1416737 [0] NCCL INFO [Service thread] Connection closed by localRank 3 cnode7-018:1407096:1407925 [0] NCCL INFO [Service thread] Connection closed by localRank 3 [cnode7-018:0/16] 2024-12-11 04:06:53,552 (average_nbest_models:70) INFO: Averaging 5best models: criterion="valid.total_count": exp_owsm/s2t_train_05b_ds_raw_bpe50000/valid.total_count.ave_5best.pth /mnt/home/williamchen/espnet/espnet2/main_funcs/average_nbest_models.py:78: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. _loaded[e] = torch.load( # Accounting: begin_time=1733822905 # Accounting: end_time=1733861226 # Accounting: time=38321 threads=1 # Finished at Wed Dec 11 04:07:06 CST 2024 with status 0