Upload checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins

checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/wandb/offline-run-20260125_170309-vlm_gym_colorization_one_img_lr2e_5_mse_only_ins-run0/files/output.log

@@ -795,49 +795,6 @@ wandb: For more information, check out the docs at: https://weave-docs.wandb.ai/
 [2026-01-25 21:54:23] (step=0000784) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-25 21:54:50] (step=0000785) Train Loss mse: 0.0068, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-25 21:55:12] (step=0000786) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:55:36] (step=0000787) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 21:55:56] (step=0000788) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:56:14] (step=0000789) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:56:38] (step=0000790) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 21:57:00] (step=0000791) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:57:17] (step=0000792) Train Loss mse: 0.0088, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-25 21:57:38] (step=0000793) Train Loss mse: 0.0084, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:58:01] (step=0000794) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 21:58:19] (step=0000795) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:58:41] (step=0000796) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:59:01] (step=0000797) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:59:22] (step=0000798) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 21:59:44] (step=0000799) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:00:06] (step=0000800) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:00:28] (step=0000801) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:00:50] (step=0000802) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:01:06] (step=0000803) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-25 22:01:27] (step=0000804) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:01:52] (step=0000805) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:02:10] (step=0000806) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:02:32] (step=0000807) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:02:52] (step=0000808) Train Loss mse: 0.0092, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:03:10] (step=0000809) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:03:31] (step=0000810) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:03:50] (step=0000811) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:04:17] (step=0000812) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:04:37] (step=0000813) Train Loss mse: 0.0079, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:05:01] (step=0000814) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:05:22] (step=0000815) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:05:42] (step=0000816) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:06:06] (step=0000817) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:06:27] (step=0000818) Train Loss mse: 0.0091, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:06:48] (step=0000819) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:07:09] (step=0000820) Train Loss mse: 0.0086, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:07:31] (step=0000821) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:07:49] (step=0000822) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-25 22:08:12] (step=0000823) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:08:30] (step=0000824) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-25 22:08:52] (step=0000825) Train Loss mse: 0.0099, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:09:13] (step=0000826) Train Loss mse: 0.0085, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:09:34] (step=0000827) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-25 22:09:58] (step=0000828) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-[2026-01-25 22:10:18] (step=0000829) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 FullyShardedDataParallel(
   (_fsdp_wrapped_module): Bagel(
     (language_model): Qwen2ForCausalLM(
@@ -1024,20 +981,49 @@ Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorizati
   fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
   fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
 ce_avg: 0.0, mse_avg: 0.007997258566319942
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step1000
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007652191445231438
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step1500
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.00800316222012043
+[2026-01-25 21:55:36] (step=0000787) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 21:55:56] (step=0000788) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:56:14] (step=0000789) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:56:38] (step=0000790) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 21:57:00] (step=0000791) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:57:17] (step=0000792) Train Loss mse: 0.0088, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-25 21:57:38] (step=0000793) Train Loss mse: 0.0084, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:58:01] (step=0000794) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 21:58:19] (step=0000795) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:58:41] (step=0000796) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:59:01] (step=0000797) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:59:22] (step=0000798) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 21:59:44] (step=0000799) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:00:06] (step=0000800) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:00:28] (step=0000801) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:00:50] (step=0000802) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:01:06] (step=0000803) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-25 22:01:27] (step=0000804) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:01:52] (step=0000805) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:02:10] (step=0000806) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:02:32] (step=0000807) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:02:52] (step=0000808) Train Loss mse: 0.0092, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:03:10] (step=0000809) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:03:31] (step=0000810) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:03:50] (step=0000811) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:04:17] (step=0000812) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:04:37] (step=0000813) Train Loss mse: 0.0079, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:05:01] (step=0000814) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:05:22] (step=0000815) Train Loss mse: 0.0078, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:05:42] (step=0000816) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:06:06] (step=0000817) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:06:27] (step=0000818) Train Loss mse: 0.0091, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:06:48] (step=0000819) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:07:09] (step=0000820) Train Loss mse: 0.0086, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:07:31] (step=0000821) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:07:49] (step=0000822) Train Loss mse: 0.0083, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-25 22:08:12] (step=0000823) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:08:30] (step=0000824) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-25 22:08:52] (step=0000825) Train Loss mse: 0.0099, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:09:13] (step=0000826) Train Loss mse: 0.0085, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:09:34] (step=0000827) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-25 22:09:58] (step=0000828) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+[2026-01-25 22:10:18] (step=0000829) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-25 22:10:41] (step=0000830) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-25 22:10:58] (step=0000831) Train Loss mse: 0.0079, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-25 22:11:18] (step=0000832) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
@@ -2053,6 +2039,20 @@ ce_avg: 0.0, mse_avg: 0.00800316222012043
 [2026-01-26 04:13:26] (step=0001842) Train Loss mse: 0.0062, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 04:13:48] (step=0001843) Train Loss mse: 0.0079, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 04:14:13] (step=0001844) Train Loss mse: 0.0065, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step1000
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007652191445231438
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step1500
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.00800316222012043
 [2026-01-26 04:14:36] (step=0001845) Train Loss mse: 0.0067, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 04:14:57] (step=0001846) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 04:15:20] (step=0001847) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
@@ -2136,20 +2136,6 @@ ce_avg: 0.0, mse_avg: 0.00800316222012043
 [2026-01-26 04:42:39] (step=0001925) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 04:43:01] (step=0001926) Train Loss mse: 0.0072, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 04:43:23] (step=0001927) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step2000
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.0081106498837471
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step2500
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007652428932487965
 [2026-01-26 04:43:45] (step=0001928) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 04:44:06] (step=0001929) Train Loss mse: 0.0069, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 04:44:28] (step=0001930) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
@@ -3082,6 +3068,20 @@ ce_avg: 0.0, mse_avg: 0.007652428932487965
 [2026-01-26 10:17:47] (step=0002857) Train Loss mse: 0.0082, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 10:18:07] (step=0002858) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 10:18:28] (step=0002859) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
+[2026-01-26 10:18:49
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step2000
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.0081106498837471
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step2500
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
 [2026-01-26 10:18:49] (step=0002860) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 10:19:12] (step=0002861) Train Loss mse: 0.0067, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 10:19:31] (step=0002862) Train Loss mse: 0.0065, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
@@ -3152,27 +3152,6 @@ ce_avg: 0.0, mse_avg: 0.007652428932487965
 [2026-01-26 10:42:06] (step=0002927) Train Loss mse: 0.0056, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 10:42:29] (step=0002928) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 10:42:50] (step=0002929) Train Loss mse: 0.0079, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
-[2026-01-26 10:43:13
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step3000
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007834003306925297
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step3500
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007766008842736483
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step4000
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
 [2026-01-26 10:43:13] (step=0002930) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 10:43:33] (step=0002931) Train Loss mse: 0.0062, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 10:43:54] (step=0002932) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
@@ -3988,6 +3967,34 @@ Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorizati
 [2026-01-26 15:35:43] (step=0003742) Train Loss mse: 0.0077, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 15:36:04] (step=0003743) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 15:36:28] (step=0003744) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step3000
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007834003306925297
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step3500
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007766008842736483
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step4000
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007558991201221943
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step4500
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007897508330643177
 [2026-01-26 15:36:52] (step=0003745) Train Loss mse: 0.0076, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 15:37:15] (step=0003746) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 15:37:35] (step=0003747) Train Loss mse: 0.0071, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
@@ -4179,20 +4186,6 @@ Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorizati
 [2026-01-26 16:44:11] (step=0003933) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 16:44:29] (step=0003934) Train Loss mse: 0.0067, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-26 16:44:52] (step=0003935) Train Loss mse: 0.0063, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step4500
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007897508330643177
-base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step5000
-Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.007832281291484833
 [2026-01-26 16:45:14] (step=0003936) Train Loss mse: 0.0067, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 16:45:36] (step=0003937) Train Loss mse: 0.0056, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 16:45:55] (step=0003938) Train Loss mse: 0.0073, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
@@ -5098,6 +5091,13 @@ ce_avg: 0.0, mse_avg: 0.007832281291484833
 [2026-01-26 22:10:09] (step=0004838) Train Loss mse: 0.0080, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 22:10:34] (step=0004839) Train Loss mse: 0.0075, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 22:10:59] (step=0004840) Train Loss mse: 0.0074, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_colorization_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_colorization_one_img_lr2e_5_mse_only_ins_step5000
+Preparing Dataset vlm_gym_colorization_mse_loss_only_evalonce/vlm_gym_colorization_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_colorization_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.007832281291484833
 [2026-01-26 22:11:23] (step=0004841) Train Loss mse: 0.0063, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,
 [2026-01-26 22:11:42] (step=0004842) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.05,
 [2026-01-26 22:12:06] (step=0004843) Train Loss mse: 0.0070, Train Loss ce: 0.0000, Train Steps/Sec: 0.04,