Upload folder using huggingface_hub

CKPTS/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/000-GPT-XL/log.txt

@@ -0,0 +1,10 @@
+[2025-04-23 09:11:27] Experiment directory created at checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/000-GPT-XL
+[2025-04-23 09:11:27] Namespace(data_path='/tmp/haozhezhao/MLLMG/jsonl_data/merged_train_set_set_subject_400k_recap_t2i_400k_flux_200k_midjourney_150k_recovery_150k_grounding_100fluxseg_50samseg.jsonl', cloud_save_path='/tmp/haozhezhao/checkpoint', no_local_save=False, vq_model='VQ-16', vq_ckpt='/tmp/haozhezhao/model/llamagen_t2i/vq_ds16_t2i.pt', codebook_size=16384, codebook_embed_dim=8, gpt_model='GPT-XL', gpt_ckpt='/tmp/haozhezhao/MLLMG_ckpkts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt', gpt_type='t2i', vocab_size=16384, cls_token_num=512, dropout_p=0.1, token_dropout_p=0.1, drop_path=0.0, no_compile=False, results_dir='checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated', dataset='ti2i', image_size=512, downsample_size=16, num_classes=1000, epochs=2, lr=0.0003, weight_decay=0.05, beta1=0.9, beta2=0.95, max_grad_norm=1.0, global_batch_size=64, global_seed=0, num_workers=8, log_every=25, ckpt_every=2000, gradient_accumulation_steps=4, mixed_precision='bf16', val_data_path='/tmp/haozhezhao/MLLMG/jsonl_data/dreambench_plus_valid.jsonl', use_vision_tower=True, model_name_or_path='/tmp/haozhezhao/model/blip2-flan-t5-xl', image_place_holder='<image>', processor_path=None, do_eval=True, max_eval_samples=500, train_text_encoder=True, no_left_padding=False, cfg_scale=7.5, top_k=16384, temperature=0.9, top_p=1.0, eval_steps=2000, project_name='llamagen_ti2i', load_from_checkpoint='/tmp/haozhezhao/MLLMG/checkpoint/FIXed_3M_1epoch_step32000.pt', warmup=0.05, lr_decay_style='cosine', lr_decay_ratio=0.1, train_iters=500000, class_dropout_prob=0.1, with_image_only=False, image_only_rate=0.1, stage2=False, subject_driven=True, load_subject_embedding=None, reference_data_path='/tmp/haozhezhao/MLLMG/cc12m_reference_tunnel.jsonl', multimodal_encoder='llava', do_recovery=True, no_replace=False, resume=False, dreambench_eval=True, find_unused_parameters=True, load_visual_encoder=False, continue_stage1=False, replace_subject=False, train_all=True, save_total_limit=1, load_language_projection='/tmp/haozhezhao/MLLMG/llava-v1.5-flant5_fixed-pretrain/mm_projector.bin', mm_vision_tower='openai/clip-vit-large-patch14', load_fixed_llamagen=True, unfreeze_output=False, fix='gpt-empty-fix', rank=0, world_size=8, gpu=0, dist_url='env://', distributed=True, dist_backend='nccl')
+[2025-04-23 09:11:27] Starting rank=0, seed=0, world_size=8.
+[2025-04-23 09:11:27] text_config is None. Initializing the text config with default values (`OPTConfig`).
+[2025-04-23 09:12:39] GPT Parameters: 2,310,680,832
+[2025-04-23 09:12:39] num decayed parameter tensors: 356, with 2,007,303,168 parameters
+[2025-04-23 09:12:39] num non-decayed parameter tensors: 124, with 197,888 parameters
+[2025-04-23 09:12:39] using fused AdamW: True
+[2025-04-23 09:12:49] Dataset contains 1,313,682 images
+[2025-04-23 09:12:49] Train iters 41052 , warmup 2052.6, len of loader 20526

CKPTS/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/001-GPT-XL/log.txt

@@ -0,0 +1,18 @@
+[2025-04-23 10:50:20] Experiment directory created at checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/001-GPT-XL
+[2025-04-23 10:50:20] Namespace(data_path='/tmp/haozhezhao/MLLMG/jsonl_data/merged_train_set_set_subject_400k_recap_t2i_400k_flux_200k_midjourney_150k_recovery_150k_grounding_100fluxseg_50samseg.jsonl', cloud_save_path='/tmp/haozhezhao/checkpoint', no_local_save=False, vq_model='VQ-16', vq_ckpt='/tmp/haozhezhao/model/llamagen_t2i/vq_ds16_t2i.pt', codebook_size=16384, codebook_embed_dim=8, gpt_model='GPT-XL', gpt_ckpt='/tmp/haozhezhao/MLLMG/MLLMG_ckpts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt', gpt_type='t2i', vocab_size=16384, cls_token_num=512, dropout_p=0.1, token_dropout_p=0.1, drop_path=0.0, no_compile=False, results_dir='checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated', dataset='ti2i', image_size=512, downsample_size=16, num_classes=1000, epochs=2, lr=0.0003, weight_decay=0.05, beta1=0.9, beta2=0.95, max_grad_norm=1.0, global_batch_size=64, global_seed=0, num_workers=8, log_every=25, ckpt_every=2000, gradient_accumulation_steps=4, mixed_precision='bf16', val_data_path='/tmp/haozhezhao/MLLMG/jsonl_data/dreambench_plus_valid.jsonl', use_vision_tower=True, model_name_or_path='/tmp/haozhezhao/model/blip2-flan-t5-xl', image_place_holder='<image>', processor_path=None, do_eval=True, max_eval_samples=500, train_text_encoder=True, no_left_padding=False, cfg_scale=7.5, top_k=16384, temperature=0.9, top_p=1.0, eval_steps=2000, project_name='llamagen_ti2i', load_from_checkpoint='/tmp/haozhezhao/MLLMG/checkpoint/FIXed_3M_1epoch_step32000.pt', warmup=0.05, lr_decay_style='cosine', lr_decay_ratio=0.1, train_iters=500000, class_dropout_prob=0.1, with_image_only=False, image_only_rate=0.1, stage2=False, subject_driven=True, load_subject_embedding=None, reference_data_path='/tmp/haozhezhao/MLLMG/cc12m_reference_tunnel.jsonl', multimodal_encoder='llava', do_recovery=True, no_replace=False, resume=False, dreambench_eval=True, find_unused_parameters=True, load_visual_encoder=False, continue_stage1=False, replace_subject=False, train_all=True, save_total_limit=1, load_language_projection='/tmp/haozhezhao/MLLMG/llava-v1.5-flant5_fixed-pretrain/mm_projector.bin', mm_vision_tower='openai/clip-vit-large-patch14', load_fixed_llamagen=True, unfreeze_output=False, fix='gpt-empty-fix', rank=0, world_size=8, gpu=0, dist_url='env://', distributed=True, dist_backend='nccl')
+[2025-04-23 10:50:20] Starting rank=0, seed=0, world_size=8.
+[2025-04-23 10:50:20] text_config is None. Initializing the text config with default values (`OPTConfig`).
+[2025-04-23 10:51:26] GPT Parameters: 2,310,680,832
+[2025-04-23 10:51:26] num decayed parameter tensors: 356, with 2,007,303,168 parameters
+[2025-04-23 10:51:26] num non-decayed parameter tensors: 124, with 197,888 parameters
+[2025-04-23 10:51:26] using fused AdamW: True
+[2025-04-23 10:51:35] Dataset contains 1,313,682 images
+[2025-04-23 10:51:35] Train iters 41052 , warmup 2052.6, len of loader 20526
+[2025-04-23 10:52:03] ### LOAD  pretraining weights from checkpoint: /tmp/haozhezhao/MLLMG/MLLMG_ckpts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt
+[2025-04-23 10:52:03] Initial state: steps=0, epochs=0
+[2025-04-23 10:52:03] compiling the model... (may take several minutes)
+[2025-04-23 10:52:04] freeze the vit
+[2025-04-23 10:52:04] ***** total param is 2310680832 *****
+[2025-04-23 10:52:04] ***** total trained param is 2007501056 *****
+[2025-04-23 10:52:06] Training for 2 epochs...
+[2025-04-23 10:52:06] Beginning epoch 0...

CKPTS/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/003-GPT-XL/log.txt

@@ -0,0 +1,4 @@
+[2025-04-23 11:01:15] Experiment directory created at checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/003-GPT-XL
+[2025-04-23 11:01:15] Namespace(data_path='/tmp/haozhezhao/MLLMG/jsonl_data/merged_train_set_set_subject_400k_recap_t2i_400k_flux_200k_midjourney_150k_recovery_150k_grounding_100fluxseg_50samseg.jsonl', cloud_save_path='/tmp/haozhezhao/checkpoint', no_local_save=False, vq_model='VQ-16', vq_ckpt='/tmp/haozhezhao/model/llamagen_t2i/vq_ds16_t2i.pt', codebook_size=16384, codebook_embed_dim=8, gpt_model='GPT-XL', gpt_ckpt='/tmp/haozhezhao/MLLMG/MLLMG_ckpts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt', gpt_type='t2i', vocab_size=16384, cls_token_num=512, dropout_p=0.1, token_dropout_p=0.1, drop_path=0.0, no_compile=False, results_dir='checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated', dataset='ti2i', image_size=512, downsample_size=16, num_classes=1000, epochs=2, lr=0.0003, weight_decay=0.05, beta1=0.9, beta2=0.95, max_grad_norm=1.0, global_batch_size=64, global_seed=0, num_workers=8, log_every=25, ckpt_every=2000, gradient_accumulation_steps=4, mixed_precision='bf16', val_data_path='/tmp/haozhezhao/MLLMG/jsonl_data/dreambench_plus_valid.jsonl', use_vision_tower=True, model_name_or_path='/tmp/haozhezhao/model/blip2-flan-t5-xl', image_place_holder='<image>', processor_path=None, do_eval=True, max_eval_samples=500, train_text_encoder=True, no_left_padding=False, cfg_scale=7.5, top_k=16384, temperature=0.9, top_p=1.0, eval_steps=2000, project_name='llamagen_ti2i', load_from_checkpoint='/tmp/haozhezhao/MLLMG/checkpoint/FIXed_3M_1epoch_step32000.pt', warmup=0.05, lr_decay_style='cosine', lr_decay_ratio=0.1, train_iters=500000, class_dropout_prob=0.1, with_image_only=False, image_only_rate=0.1, stage2=False, subject_driven=True, load_subject_embedding=None, reference_data_path='/tmp/haozhezhao/MLLMG/cc12m_reference_tunnel.jsonl', multimodal_encoder='llava', do_recovery=True, no_replace=False, resume=False, dreambench_eval=True, find_unused_parameters=True, load_visual_encoder=False, continue_stage1=False, replace_subject=False, train_all=True, save_total_limit=1, load_language_projection='/tmp/haozhezhao/MLLMG/llava-v1.5-flant5_fixed-pretrain/mm_projector.bin', mm_vision_tower='openai/clip-vit-large-patch14', load_fixed_llamagen=True, unfreeze_output=False, fix='gpt-empty-fix', rank=0, world_size=8, gpu=0, dist_url='env://', distributed=True, dist_backend='nccl')
+[2025-04-23 11:01:15] Starting rank=0, seed=0, world_size=8.
+[2025-04-23 11:01:15] text_config is None. Initializing the text config with default values (`OPTConfig`).

CKPTS/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/004-GPT-XL/log.txt

@@ -0,0 +1,99 @@
+[2025-04-23 11:12:04] Experiment directory created at checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/004-GPT-XL
+[2025-04-23 11:12:04] Namespace(data_path='/tmp/haozhezhao/MLLMG/jsonl_data/merged_train_set_set_subject_400k_recap_t2i_400k_flux_200k_midjourney_150k_recovery_150k_grounding_100fluxseg_50samseg.jsonl', cloud_save_path='/tmp/haozhezhao/checkpoint', no_local_save=False, vq_model='VQ-16', vq_ckpt='/tmp/haozhezhao/model/llamagen_t2i/vq_ds16_t2i.pt', codebook_size=16384, codebook_embed_dim=8, gpt_model='GPT-XL', gpt_ckpt='/tmp/haozhezhao/MLLMG/MLLMG_ckpts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt', gpt_type='t2i', vocab_size=16384, cls_token_num=512, dropout_p=0.1, token_dropout_p=0.1, drop_path=0.0, no_compile=False, results_dir='checkpoint/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated', dataset='ti2i', image_size=512, downsample_size=16, num_classes=1000, epochs=2, lr=0.0003, weight_decay=0.05, beta1=0.9, beta2=0.95, max_grad_norm=1.0, global_batch_size=56, global_seed=0, num_workers=8, log_every=25, ckpt_every=2000, gradient_accumulation_steps=4, mixed_precision='bf16', val_data_path='/tmp/haozhezhao/MLLMG/jsonl_data/dreambench_plus_valid.jsonl', use_vision_tower=True, model_name_or_path='/tmp/haozhezhao/model/blip2-flan-t5-xl', image_place_holder='<image>', processor_path=None, do_eval=True, max_eval_samples=500, train_text_encoder=True, no_left_padding=False, cfg_scale=7.5, top_k=16384, temperature=0.9, top_p=1.0, eval_steps=2000, project_name='llamagen_ti2i', load_from_checkpoint='/tmp/haozhezhao/MLLMG/checkpoint/FIXed_3M_1epoch_step32000.pt', warmup=0.05, lr_decay_style='cosine', lr_decay_ratio=0.1, train_iters=500000, class_dropout_prob=0.1, with_image_only=False, image_only_rate=0.1, stage2=False, subject_driven=True, load_subject_embedding=None, reference_data_path='/tmp/haozhezhao/MLLMG/cc12m_reference_tunnel.jsonl', multimodal_encoder='llava', do_recovery=True, no_replace=False, resume=False, dreambench_eval=True, find_unused_parameters=True, load_visual_encoder=False, continue_stage1=False, replace_subject=False, train_all=True, save_total_limit=1, load_language_projection='/tmp/haozhezhao/MLLMG/llava-v1.5-flant5_fixed-pretrain/mm_projector.bin', mm_vision_tower='openai/clip-vit-large-patch14', load_fixed_llamagen=True, unfreeze_output=False, fix='gpt-empty-fix', rank=0, world_size=8, gpu=0, dist_url='env://', distributed=True, dist_backend='nccl')
+[2025-04-23 11:12:04] Starting rank=0, seed=0, world_size=8.
+[2025-04-23 11:12:04] text_config is None. Initializing the text config with default values (`OPTConfig`).
+[2025-04-23 11:13:13] GPT Parameters: 2,310,680,832
+[2025-04-23 11:13:13] num decayed parameter tensors: 356, with 2,007,303,168 parameters
+[2025-04-23 11:13:13] num non-decayed parameter tensors: 124, with 197,888 parameters
+[2025-04-23 11:13:13] using fused AdamW: True
+[2025-04-23 11:13:22] Dataset contains 1,313,682 images
+[2025-04-23 11:13:22] Train iters 46916 , warmup 2345.8, len of loader 23458
+[2025-04-23 11:13:38] ### LOAD  pretraining weights from checkpoint: /tmp/haozhezhao/MLLMG/MLLMG_ckpts/checkpoints/AfterI2I_llavat5_subject_train_all_T5_subject_t2i_ti2i_120_w_flux_segment_noimageonly_empty_fix_train_all/0100000.pt
+[2025-04-23 11:13:38] Initial state: steps=0, epochs=0
+[2025-04-23 11:13:38] compiling the model... (may take several minutes)
+[2025-04-23 11:13:39] freeze the vit
+[2025-04-23 11:13:39] ***** total param is 2310680832 *****
+[2025-04-23 11:13:39] ***** total trained param is 2007501056 *****
+[2025-04-23 11:13:40] Training for 2 epochs...
+[2025-04-23 11:13:40] Beginning epoch 0...
+[2025-04-23 11:18:40] (step=0000025) Train Loss: 6.6005, Train Steps/Sec: 0.08
+[2025-04-23 11:19:29] (step=0000050) Train Loss: 6.6556, Train Steps/Sec: 0.51
+[2025-04-23 11:21:08] (step=0000075) Train Loss: 6.4599, Train Steps/Sec: 0.25
+[2025-04-23 11:22:11] (step=0000100) Train Loss: 6.3558, Train Steps/Sec: 0.40
+[2025-04-23 11:22:33] (step=0000125) Train Loss: 6.2557, Train Steps/Sec: 1.17
+[2025-04-23 11:23:23] (step=0000150) Train Loss: 6.1054, Train Steps/Sec: 0.50
+[2025-04-23 11:24:13] (step=0000175) Train Loss: 6.1218, Train Steps/Sec: 0.50
+[2025-04-23 11:24:35] (step=0000200) Train Loss: 6.0449, Train Steps/Sec: 1.15
+[2025-04-23 11:24:56] (step=0000225) Train Loss: 5.9306, Train Steps/Sec: 1.17
+[2025-04-23 11:25:46] (step=0000250) Train Loss: 5.8833, Train Steps/Sec: 0.50
+[2025-04-23 11:27:04] (step=0000275) Train Loss: 5.9763, Train Steps/Sec: 0.32
+[2025-04-23 11:27:25] (step=0000300) Train Loss: 5.8122, Train Steps/Sec: 1.17
+[2025-04-23 11:27:47] (step=0000325) Train Loss: 5.9543, Train Steps/Sec: 1.17
+[2025-04-23 11:28:08] (step=0000350) Train Loss: 5.8445, Train Steps/Sec: 1.17
+[2025-04-23 11:28:29] (step=0000375) Train Loss: 5.8658, Train Steps/Sec: 1.17
+[2025-04-23 11:29:20] (step=0000400) Train Loss: 5.7741, Train Steps/Sec: 0.49
+[2025-04-23 11:29:41] (step=0000425) Train Loss: 5.8472, Train Steps/Sec: 1.17
+[2025-04-23 11:30:03] (step=0000450) Train Loss: 5.8283, Train Steps/Sec: 1.17
+[2025-04-23 11:30:25] (step=0000475) Train Loss: 5.8059, Train Steps/Sec: 1.12
+[2025-04-23 11:30:46] (step=0000500) Train Loss: 5.8193, Train Steps/Sec: 1.17
+[2025-04-23 11:31:08] (step=0000525) Train Loss: 5.8235, Train Steps/Sec: 1.17
+[2025-04-23 11:31:29] (step=0000550) Train Loss: 5.8375, Train Steps/Sec: 1.17
+[2025-04-23 11:31:50] (step=0000575) Train Loss: 5.8427, Train Steps/Sec: 1.17
+[2025-04-23 11:32:12] (step=0000600) Train Loss: 5.7903, Train Steps/Sec: 1.16
+[2025-04-23 11:32:33] (step=0000625) Train Loss: 5.8660, Train Steps/Sec: 1.17
+[2025-04-23 11:32:55] (step=0000650) Train Loss: 5.7774, Train Steps/Sec: 1.17
+[2025-04-23 11:33:16] (step=0000675) Train Loss: 5.7929, Train Steps/Sec: 1.17
+[2025-04-23 11:33:37] (step=0000700) Train Loss: 5.9227, Train Steps/Sec: 1.17
+[2025-04-23 11:33:59] (step=0000725) Train Loss: 5.7613, Train Steps/Sec: 1.17
+[2025-04-23 11:34:21] (step=0000750) Train Loss: 5.8464, Train Steps/Sec: 1.11
+[2025-04-23 11:34:42] (step=0000775) Train Loss: 5.8326, Train Steps/Sec: 1.17
+[2025-04-23 11:35:05] (step=0000800) Train Loss: 5.7977, Train Steps/Sec: 1.11
+[2025-04-23 11:35:26] (step=0000825) Train Loss: 5.7656, Train Steps/Sec: 1.17
+[2025-04-23 11:35:48] (step=0000850) Train Loss: 5.7243, Train Steps/Sec: 1.17
+[2025-04-23 11:36:11] (step=0000875) Train Loss: 5.7829, Train Steps/Sec: 1.07
+[2025-04-23 11:36:34] (step=0000900) Train Loss: 5.7715, Train Steps/Sec: 1.10
+[2025-04-23 11:36:55] (step=0000925) Train Loss: 5.8266, Train Steps/Sec: 1.17
+[2025-04-23 11:37:16] (step=0000950) Train Loss: 5.8739, Train Steps/Sec: 1.17
+[2025-04-23 11:37:37] (step=0000975) Train Loss: 5.7752, Train Steps/Sec: 1.17
+[2025-04-23 11:37:59] (step=0001000) Train Loss: 5.7676, Train Steps/Sec: 1.17
+[2025-04-23 11:38:22] (step=0001025) Train Loss: 5.7233, Train Steps/Sec: 1.06
+[2025-04-23 11:38:44] (step=0001050) Train Loss: 5.7606, Train Steps/Sec: 1.17
+[2025-04-23 11:39:05] (step=0001075) Train Loss: 5.7211, Train Steps/Sec: 1.17
+[2025-04-23 11:39:26] (step=0001100) Train Loss: 5.7764, Train Steps/Sec: 1.17
+[2025-04-23 11:39:48] (step=0001125) Train Loss: 5.7317, Train Steps/Sec: 1.18
+[2025-04-23 11:40:10] (step=0001150) Train Loss: 5.7573, Train Steps/Sec: 1.12
+[2025-04-23 11:40:31] (step=0001175) Train Loss: 5.7894, Train Steps/Sec: 1.17
+[2025-04-23 11:40:53] (step=0001200) Train Loss: 5.7181, Train Steps/Sec: 1.17
+[2025-04-23 11:41:14] (step=0001225) Train Loss: 5.7548, Train Steps/Sec: 1.17
+[2025-04-23 11:41:35] (step=0001250) Train Loss: 5.6894, Train Steps/Sec: 1.17
+[2025-04-23 11:41:57] (step=0001275) Train Loss: 5.7133, Train Steps/Sec: 1.17
+[2025-04-23 11:42:18] (step=0001300) Train Loss: 5.7372, Train Steps/Sec: 1.17
+[2025-04-23 11:42:39] (step=0001325) Train Loss: 5.8133, Train Steps/Sec: 1.17
+[2025-04-23 11:43:01] (step=0001350) Train Loss: 5.7414, Train Steps/Sec: 1.17
+[2025-04-23 11:43:22] (step=0001375) Train Loss: 5.7914, Train Steps/Sec: 1.17
+[2025-04-23 11:43:44] (step=0001400) Train Loss: 5.6661, Train Steps/Sec: 1.17
+[2025-04-23 11:44:05] (step=0001425) Train Loss: 5.7186, Train Steps/Sec: 1.17
+[2025-04-23 11:44:26] (step=0001450) Train Loss: 5.8405, Train Steps/Sec: 1.17
+[2025-04-23 11:44:48] (step=0001475) Train Loss: 5.7429, Train Steps/Sec: 1.17
+[2025-04-23 11:45:09] (step=0001500) Train Loss: 5.6897, Train Steps/Sec: 1.17
+[2025-04-23 11:45:30] (step=0001525) Train Loss: 5.7404, Train Steps/Sec: 1.17
+[2025-04-23 11:45:53] (step=0001550) Train Loss: 5.7304, Train Steps/Sec: 1.12
+[2025-04-23 11:46:14] (step=0001575) Train Loss: 5.7393, Train Steps/Sec: 1.17
+[2025-04-23 11:46:35] (step=0001600) Train Loss: 5.7214, Train Steps/Sec: 1.17
+[2025-04-23 11:46:58] (step=0001625) Train Loss: 5.7278, Train Steps/Sec: 1.12
+[2025-04-23 11:47:21] (step=0001650) Train Loss: 5.7655, Train Steps/Sec: 1.08
+[2025-04-23 11:47:42] (step=0001675) Train Loss: 5.7132, Train Steps/Sec: 1.17
+[2025-04-23 11:48:05] (step=0001700) Train Loss: 5.8445, Train Steps/Sec: 1.11
+[2025-04-23 11:48:26] (step=0001725) Train Loss: 5.7540, Train Steps/Sec: 1.17
+[2025-04-23 11:48:48] (step=0001750) Train Loss: 5.6806, Train Steps/Sec: 1.17
+[2025-04-23 11:49:10] (step=0001775) Train Loss: 5.7114, Train Steps/Sec: 1.11
+[2025-04-23 11:49:32] (step=0001800) Train Loss: 5.7751, Train Steps/Sec: 1.17
+[2025-04-23 11:49:53] (step=0001825) Train Loss: 5.7199, Train Steps/Sec: 1.17
+[2025-04-23 11:50:16] (step=0001850) Train Loss: 5.7680, Train Steps/Sec: 1.07
+[2025-04-23 11:50:38] (step=0001875) Train Loss: 5.7145, Train Steps/Sec: 1.17
+[2025-04-23 11:50:59] (step=0001900) Train Loss: 5.6735, Train Steps/Sec: 1.17
+[2025-04-23 11:51:20] (step=0001925) Train Loss: 5.8258, Train Steps/Sec: 1.17
+[2025-04-23 11:51:42] (step=0001950) Train Loss: 5.7326, Train Steps/Sec: 1.17
+[2025-04-23 11:52:03] (step=0001975) Train Loss: 5.6651, Train Steps/Sec: 1.17
+[2025-04-23 11:52:25] (step=0002000) Train Loss: 5.6799, Train Steps/Sec: 1.17
+[2025-04-23 11:52:25] text_config is None. Initializing the text config with default values (`OPTConfig`).

CKPTS/Afteri2i_100kckpt_LLAVAT5_empty_Fixed_just_segment_pretrain_stage3_dreambench_recap_400k_t2i_400k_flux_200k_midjourney_150k_recovery_150k_3e4_epoch2_half_rotated/005-GPT-XL/checkpoints/0046000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d8a592e16377c249d4988c125dbe55d5d04d491fd305c8184ebb29e85d2eafc
+size 17322479662