Upload checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins

checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/wandb/offline-run-20260129_221543-vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins-run0/files/output.log

@@ -850,6 +850,15 @@ wandb: For more information, check out the docs at: https://weave-docs.wandb.ai/
 [2026-01-30 02:20:25] (step=0000839) Train Loss mse: 0.0258, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 02:20:41] (step=0000840) Train Loss mse: 0.0242, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 02:20:58] (step=0000841) Train Loss mse: 0.0250, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:21:15] (step=0000842) Train Loss mse: 0.0278, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:21:32] (step=0000843) Train Loss mse: 0.0222, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:21:49] (step=0000844) Train Loss mse: 0.0243, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:22:05] (step=0000845) Train Loss mse: 0.0232, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:22:22] (step=0000846) Train Loss mse: 0.0242, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:22:39] (step=0000847) Train Loss mse: 0.0240, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:22:56] (step=0000848) Train Loss mse: 0.0224, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:23:12] (step=0000849) Train Loss mse: 0.0229, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 02:23:29] (step=0000850) Train Loss mse: 0.0265, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 FullyShardedDataParallel(
   (_fsdp_wrapped_module): Bagel(
     (language_model): Qwen2ForCausalLM(
@@ -1043,15 +1052,13 @@ Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce
   fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
   fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
 ce_avg: 0.0, mse_avg: 0.024334488436579704
-[2026-01-30 02:21:15] (step=0000842) Train Loss mse: 0.0278, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:21:32] (step=0000843) Train Loss mse: 0.0222, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:21:49] (step=0000844) Train Loss mse: 0.0243, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:22:05] (step=0000845) Train Loss mse: 0.0232, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:22:22] (step=0000846) Train Loss mse: 0.0242, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:22:39] (step=0000847) Train Loss mse: 0.0240, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:22:56] (step=0000848) Train Loss mse: 0.0224, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:23:12] (step=0000849) Train Loss mse: 0.0229, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 02:23:29] (step=0000850) Train Loss mse: 0.0265, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step2000
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024451851844787598
 [2026-01-30 02:23:46] (step=0000851) Train Loss mse: 0.0230, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 02:24:03] (step=0000852) Train Loss mse: 0.0253, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 02:24:20] (step=0000853) Train Loss mse: 0.0239, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -2200,6 +2207,20 @@ ce_avg: 0.0, mse_avg: 0.024334488436579704
 [2026-01-30 07:47:18] (step=0001996) Train Loss mse: 0.0242, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 07:47:34] (step=0001997) Train Loss mse: 0.0238, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 07:47:51] (step=0001998) Train Loss mse: 0.0238, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step2500
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024314723908901215
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step3000
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024401195347309113
 [2026-01-30 07:48:08] (step=0001999) Train Loss mse: 0.0235, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 07:49:58] (step=0002000) Train Loss mse: 0.0214, Train Loss ce: 0.0000, Train Steps/Sec: 0.01,
 [2026-01-30 07:50:15] (step=0002001) Train Loss mse: 0.0226, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -2264,20 +2285,6 @@ ce_avg: 0.0, mse_avg: 0.024334488436579704
 [2026-01-30 08:06:43] (step=0002060) Train Loss mse: 0.0223, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 08:06:59] (step=0002061) Train Loss mse: 0.0213, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 08:07:16] (step=0002062) Train Loss mse: 0.0225, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step2000
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024451851844787598
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step2500
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024314723908901215
 [2026-01-30 08:07:33] (step=0002063) Train Loss mse: 0.0232, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 08:07:50] (step=0002064) Train Loss mse: 0.0223, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 08:08:07] (step=0002065) Train Loss mse: 0.0223, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -3322,28 +3329,34 @@ ce_avg: 0.0, mse_avg: 0.024314723908901215
 [2026-01-30 13:01:56] (step=0003104) Train Loss mse: 0.0206, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:02:13] (step=0003105) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:02:30] (step=0003106) Train Loss mse: 0.0204, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step3000
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024401195347309113
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step3500
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024277793243527412
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step4000
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024172412231564522
- 2026-01-30 13:10:20] (step=0003134) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:02:47] (step=0003107) Train Loss mse: 0.0242, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:03:04] (step=0003108) Train Loss mse: 0.0230, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:03:20] (step=0003109) Train Loss mse: 0.0224, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:03:37] (step=0003110) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:03:54] (step=0003111) Train Loss mse: 0.0227, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:04:11] (step=0003112) Train Loss mse: 0.0232, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:04:28] (step=0003113) Train Loss mse: 0.0216, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:04:44] (step=0003114) Train Loss mse: 0.0245, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:05:01] (step=0003115) Train Loss mse: 0.0215, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:05:18] (step=0003116) Train Loss mse: 0.0194, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:05:35] (step=0003117) Train Loss mse: 0.0217, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:05:51] (step=0003118) Train Loss mse: 0.0233, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:06:08] (step=0003119) Train Loss mse: 0.0210, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:06:25] (step=0003120) Train Loss mse: 0.0239, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:06:42] (step=0003121) Train Loss mse: 0.0241, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:06:59] (step=0003122) Train Loss mse: 0.0211, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:07:16] (step=0003123) Train Loss mse: 0.0229, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:07:33] (step=0003124) Train Loss mse: 0.0223, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:07:50] (step=0003125) Train Loss mse: 0.0215, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:08:06] (step=0003126) Train Loss mse: 0.0209, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:08:23] (step=0003127) Train Loss mse: 0.0225, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:08:40] (step=0003128) Train Loss mse: 0.0219, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:08:57] (step=0003129) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:09:13] (step=0003130) Train Loss mse: 0.0230, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:09:30] (step=0003131) Train Loss mse: 0.0237, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:09:47] (step=0003132) Train Loss mse: 0.0203, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:10:04] (step=0003133) Train Loss mse: 0.0233, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+[2026-01-30 13:10:20] (step=0003134) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:10:37] (step=0003135) Train Loss mse: 0.0237, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:10:54] (step=0003136) Train Loss mse: 0.0221, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:11:11] (step=0003137) Train Loss mse: 0.0219, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -3364,6 +3377,20 @@ ce_avg: 0.0, mse_avg: 0.024172412231564522
 [2026-01-30 13:15:23] (step=0003152) Train Loss mse: 0.0230, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:15:40] (step=0003153) Train Loss mse: 0.0212, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:15:56] (step=0003154) Train Loss mse: 0.0249, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step3500
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024277793243527412
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step4000
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024172412231564522
 [2026-01-30 13:16:13] (step=0003155) Train Loss mse: 0.0206, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:16:30] (step=0003156) Train Loss mse: 0.0216, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 13:16:47] (step=0003157) Train Loss mse: 0.0233, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -4389,20 +4416,6 @@ ce_avg: 0.0, mse_avg: 0.024172412231564522
 [2026-01-30 18:05:06] (step=0004177) Train Loss mse: 0.0218, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:05:22] (step=0004178) Train Loss mse: 0.0224, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:05:39] (step=0004179) Train Loss mse: 0.0202, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step4500
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.02414196915924549
-base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step5000
-Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
-[eval debug] first 3 batch fingerprints:
-  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
-ce_avg: 0.0, mse_avg: 0.024185948073863983
 [2026-01-30 18:05:56] (step=0004180) Train Loss mse: 0.0244, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:06:13] (step=0004181) Train Loss mse: 0.0221, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:06:30] (step=0004182) Train Loss mse: 0.0214, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
@@ -4420,7 +4433,21 @@ ce_avg: 0.0, mse_avg: 0.024185948073863983
 [2026-01-30 18:09:51] (step=0004194) Train Loss mse: 0.0225, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:10:08] (step=0004195) Train Loss mse: 0.0217, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:10:25] (step=0004196) Train Loss mse: 0.0208, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
-[2026-01-30 18:10:41] (step=0004197) Train Loss mse: 0.0209, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step4500
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.02414196915924549
+base_dir is /dev/shm/models/checkpoints_vlm_gym_mental_rotation_3d_pad3_by_axis_one_image_lr2e_5_mse_only_ins/eval_used_rows, step_tag is vlm_gym_mental_rotation_3d_pad3_by_axis_one_img_lr2e_5_mse_only_ins_step5000
+Preparing Dataset vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce/vlm_gym_mental_rotation_3d_pad3_by_axis_val
+[eval debug] first 3 batch fingerprints:
+  fp[0]: [{'data_indexes': [0], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[1]: [{'data_indexes': [8], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+  fp[2]: [{'data_indexes': [16], 'worker_id': 0, 'dataset_name': 'vlm_gym_mental_rotation_3d_pad3_by_axis_mse_loss_only_evalonce'}]
+ce_avg: 0.0, mse_avg: 0.024185948073863983
+ 2026-01-30 18:10:41] (step=0004197) Train Loss mse: 0.0209, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:10:58] (step=0004198) Train Loss mse: 0.0209, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:11:15] (step=0004199) Train Loss mse: 0.0245, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,
 [2026-01-30 18:11:32] (step=0004200) Train Loss mse: 0.0206, Train Loss ce: 0.0000, Train Steps/Sec: 0.06,