Whalswp commited on
Commit
c6502e1
·
verified ·
1 Parent(s): 983ecce

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. Abs_6D/DiT_Layer/16/Baseline/baseline_gpu1_rolling5k_20260723_022922.log +198 -0
  2. Abs_6D/DiT_Layer/16/Baseline/baseline_gpu1_rolling5k_20260723_023110_newwandb.log +0 -0
  3. Abs_6D/DiT_Layer/16/Baseline/wandb/debug-internal.log +0 -0
  4. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/action_space_manifest.json +144 -0
  5. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/phase_schedule.json +32 -0
  6. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/resolved_config.yaml +35 -0
  7. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/source_config.yaml +63 -0
  8. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/training_state/optimizer_param_groups.json +361 -0
  9. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/training_state/training_step.json +5 -0
  10. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/action_space_manifest.json +144 -0
  11. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/phase_schedule.json +32 -0
  12. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/action_space_manifest.json +144 -0
  13. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/config.json +118 -0
  14. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/phase_schedule.json +32 -0
  15. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/prompt_manifest.json +40 -0
  16. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/train_config.json +269 -0
  17. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/prompt_manifest.json +40 -0
  18. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/resolved_config.yaml +35 -0
  19. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/source_config.yaml +63 -0
  20. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/training_state/optimizer_param_groups.json +361 -0
  21. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/rscl_gpu1_rolling5k_20260723_022922.log +198 -0
  22. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/rscl_gpu1_rolling5k_20260723_023110_newwandb.log +0 -0
  23. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/source_config.yaml +2 -0
  24. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/debug-internal.log +0 -0
  25. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/debug.log +65 -20
  26. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/config.yaml +334 -0
  27. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/output.log +44 -1
  28. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/wandb-summary.json +1 -0
  29. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug-core.log +36 -0
  30. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug-internal.log +106 -0
  31. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug.log +45 -0
  32. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/files/wandb-summary.json +1 -0
  33. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug-core.log +30 -0
  34. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug-internal.log +15 -0
  35. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug.log +13 -0
  36. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/run-cemf0rge.wandb +0 -0
  37. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/output.log +149 -0
  38. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/requirements.txt +117 -0
  39. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/wandb-metadata.json +53 -0
  40. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug-core.log +14 -0
  41. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug-internal.log +0 -0
  42. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug.log +19 -0
  43. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/config.yaml +329 -0
  44. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/output.log +96 -0
  45. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/requirements.txt +118 -0
  46. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/wandb-metadata.json +46 -0
  47. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/wandb-summary.json +1 -0
  48. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-core.log +71 -0
  49. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-internal.log +0 -0
  50. Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug.log +65 -0
Abs_6D/DiT_Layer/16/Baseline/baseline_gpu1_rolling5k_20260723_022922.log ADDED
@@ -0,0 +1,198 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ INFO 2026-07-23 02:29:36 ot_train.py:281 {'batch_size': 64,
2
+ 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/Baseline/checkpoints/030000',
3
+ 'cudnn_deterministic': False,
4
+ 'dataset': {'depth_output_unit': 'mm',
5
+ 'episodes': None,
6
+ 'eval_split': 0.0,
7
+ 'image_transforms': {'enable': False,
8
+ 'max_num_transforms': 3,
9
+ 'random_order': False,
10
+ 'tfs': {'affine': {'kwargs': {'degrees': [-5.0,
11
+ 5.0],
12
+ 'translate': [0.05,
13
+ 0.05]},
14
+ 'type': 'RandomAffine',
15
+ 'weight': 1.0},
16
+ 'brightness': {'kwargs': {'brightness': [0.8,
17
+ 1.2]},
18
+ 'type': 'ColorJitter',
19
+ 'weight': 1.0},
20
+ 'contrast': {'kwargs': {'contrast': [0.8,
21
+ 1.2]},
22
+ 'type': 'ColorJitter',
23
+ 'weight': 1.0},
24
+ 'hue': {'kwargs': {'hue': [-0.05,
25
+ 0.05]},
26
+ 'type': 'ColorJitter',
27
+ 'weight': 1.0},
28
+ 'saturation': {'kwargs': {'saturation': [0.5,
29
+ 1.5]},
30
+ 'type': 'ColorJitter',
31
+ 'weight': 1.0},
32
+ 'sharpness': {'kwargs': {'sharpness': [0.5,
33
+ 1.5]},
34
+ 'type': 'SharpnessJitter',
35
+ 'weight': 1.0}}},
36
+ 'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2',
37
+ 'return_uint8': False,
38
+ 'revision': None,
39
+ 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2',
40
+ 'streaming': False,
41
+ 'use_imagenet_stats': True,
42
+ 'video_backend': 'torchcodec'},
43
+ 'env': None,
44
+ 'env_eval_freq': 20000,
45
+ 'eval': {'batch_size': 50,
46
+ 'n_episodes': 50,
47
+ 'recording': False,
48
+ 'recording_private': False,
49
+ 'recording_repo_id': None,
50
+ 'use_async_envs': True},
51
+ 'eval_steps': 0,
52
+ 'job': {'detach': False,
53
+ 'image': 'huggingface/lerobot-gpu:latest',
54
+ 'tags': [],
55
+ 'target': None,
56
+ 'timeout': '2d'},
57
+ 'job_name': 'INSIGHT_6D_DiT16_Baseline',
58
+ 'log_freq': 200,
59
+ 'max_eval_samples': 0,
60
+ 'num_workers': 4,
61
+ 'optimizer': {'betas': [0.9, 0.999],
62
+ 'eps': 1e-08,
63
+ 'grad_clip_norm': 1.0,
64
+ 'lr': 0.0001,
65
+ 'type': 'adamw',
66
+ 'weight_decay': 1e-05},
67
+ 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/Baseline',
68
+ 'peft': None,
69
+ 'persistent_workers': True,
70
+ 'policy': {'action_decode_transform': None,
71
+ 'action_expert_num_layers': 16,
72
+ 'balance_dataset_weights': True,
73
+ 'balance_trajectory_weights': True,
74
+ 'base_model_path': 'nvidia/GR00T-N1.7-3B',
75
+ 'batch_size': 32,
76
+ 'chunk_size': 16,
77
+ 'dataloader_num_workers': 8,
78
+ 'dataset_paths': None,
79
+ 'device': 'cuda',
80
+ 'embodiment_tag': 'new_embodiment',
81
+ 'image_size': [256, 256],
82
+ 'input_features': {'observation.images.guide': {'shape': [3,
83
+ 224,
84
+ 224],
85
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
86
+ 'observation.images.right_shoulder': {'shape': [3,
87
+ 224,
88
+ 224],
89
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
90
+ 'observation.images.wrist': {'shape': [3,
91
+ 224,
92
+ 224],
93
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
94
+ 'observation.state': {'shape': [16],
95
+ 'type': <FeatureType.STATE: 'STATE'>}},
96
+ 'license': None,
97
+ 'lora_alpha': 16,
98
+ 'lora_dropout': 0.1,
99
+ 'lora_full_model': False,
100
+ 'lora_rank': 0,
101
+ 'max_action_dim': 132,
102
+ 'max_state_dim': 132,
103
+ 'max_steps': 10000,
104
+ 'model_params_fp32': True,
105
+ 'n_action_steps': 16,
106
+ 'n_obs_steps': 1,
107
+ 'normalization_mapping': {'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>,
108
+ 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>,
109
+ 'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>},
110
+ 'num_inference_timesteps': None,
111
+ 'optimizer_betas': [0.9, 0.999],
112
+ 'optimizer_eps': 1e-08,
113
+ 'optimizer_lr': 0.0001,
114
+ 'optimizer_weight_decay': 1e-05,
115
+ 'output_dir': './tmp/gr00t',
116
+ 'output_features': {'action': {'shape': [10],
117
+ 'type': <FeatureType.ACTION: 'ACTION'>}},
118
+ 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/Baseline/checkpoints/030000/pretrained_model',
119
+ 'pretrained_revision': None,
120
+ 'private': None,
121
+ 'push_to_hub': False,
122
+ 'relative_exclude_joints': [],
123
+ 'repo_id': None,
124
+ 'report_to': 'wandb',
125
+ 'resume': False,
126
+ 'rkd_angle_weight': 1.0,
127
+ 'rkd_distance_weight': 1.0,
128
+ 'rkd_enabled': False,
129
+ 'rkd_fm_loss_weight': 1.0,
130
+ 'rkd_loss_weight': 0.1,
131
+ 'rkd_student_attention_hidden_dim': 512,
132
+ 'rkd_student_attention_output_dim': 512,
133
+ 'rkd_student_pooling': 'attention',
134
+ 'rkd_vae_checkpoint': None,
135
+ 'rtc_ramp_rate': None,
136
+ 'save_steps': 1000,
137
+ 'tags': None,
138
+ 'tokenizer_assets_repo': None,
139
+ 'tune_diffusion_model': True,
140
+ 'tune_llm': False,
141
+ 'tune_projector': True,
142
+ 'tune_top_llm_layers': 0,
143
+ 'tune_visual': False,
144
+ 'tune_vlln': True,
145
+ 'type': 'groot_rkd',
146
+ 'use_amp': False,
147
+ 'use_bf16': True,
148
+ 'use_flash_attention': False,
149
+ 'use_peft': False,
150
+ 'use_relative_actions': False,
151
+ 'video_backend': 'decord',
152
+ 'warmup_ratio': 0.05,
153
+ 'warmup_steps': None},
154
+ 'prefetch_factor': 4,
155
+ 'rename_map': {},
156
+ 'resume': True,
157
+ 'reward_model': None,
158
+ 'sample_weighting': None,
159
+ 'save_checkpoint': True,
160
+ 'save_checkpoint_to_hub': False,
161
+ 'save_freq': 20000,
162
+ 'scheduler': {'name': 'cosine', 'num_warmup_steps': 500, 'type': 'diffuser'},
163
+ 'seed': 42,
164
+ 'steps': 60000,
165
+ 'tolerance_s': 0.0001,
166
+ 'use_policy_training_preset': True,
167
+ 'wandb': {'add_tags': True,
168
+ 'disable_artifact': True,
169
+ 'enable': True,
170
+ 'entity': None,
171
+ 'mode': None,
172
+ 'notes': None,
173
+ 'project': 'lerobot',
174
+ 'run_id': 'eb0sb786'}}
175
+ Traceback (most recent call last):
176
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1613, in <module>
177
+ main()
178
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1609, in main
179
+ train_module.main()
180
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 822, in main
181
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/configs/parser.py", line 320, in wrapper_inner
182
+ response = fn(cfg, *args, **kwargs)
183
+ ^^^^^^^^^^^^^^^^^^^^^^^^
184
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 285, in train
185
+ if not policy.is_trainable:
186
+ ^^^^^^^^^^^^^
187
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/common/wandb_utils.py", line 99, in __init__
188
+ wandb.init(
189
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 1528, in init
190
+ get_sentry().reraise(e)
191
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/analytics/sentry.py", line 191, in reraise
192
+ raise exc.with_traceback(tb)
193
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 1511, in init
194
+ run = wi.init(run_settings, run_config, run_printer)
195
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
196
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 961, in init
197
+ raise error
198
+ wandb.errors.errors.UsageError: You provided an invalid value for the `resume` argument. The value 'must' is not a valid option for resuming the run (eb0sb786) that has not been initialized. Please check your inputs and try again with a valid run ID. If you are trying to start a new run, please omit the `resume` argument or use `resume='allow'`.
Abs_6D/DiT_Layer/16/Baseline/baseline_gpu1_rolling5k_20260723_023110_newwandb.log ADDED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/Baseline/wandb/debug-internal.log CHANGED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/action_space_manifest.json ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 2,
3
+ "action_space": "ee_abs_rot6d",
4
+ "action_views": [
5
+ "ee_abs_rot6d"
6
+ ],
7
+ "source_feature": "action.ee_abs_rot6d",
8
+ "canonical_feature": "action",
9
+ "dtype": "float32",
10
+ "shape": [
11
+ 10
12
+ ],
13
+ "names": [
14
+ "x",
15
+ "y",
16
+ "z",
17
+ "r_col0_x",
18
+ "r_col0_y",
19
+ "r_col0_z",
20
+ "r_col1_x",
21
+ "r_col1_y",
22
+ "r_col1_z",
23
+ "gripper"
24
+ ],
25
+ "dataset_repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
26
+ "dataset_revision": "v3.0",
27
+ "dataset_total_frames": 218367,
28
+ "dataset_total_episodes": 4930,
29
+ "dataset_fps": 10,
30
+ "stats_sha256": "71698c7337070aa8267df5aea7ec08bc00d3925510f2649bc20c9201eeb669e4",
31
+ "stats": {
32
+ "min": [
33
+ -0.21784618496894836,
34
+ -0.4474811851978302,
35
+ 0.1702737659215927,
36
+ -0.9999986290931702,
37
+ -0.9999991655349731,
38
+ -0.9999999403953552,
39
+ -0.9999990463256836,
40
+ -0.9999998211860657,
41
+ -0.9999988079071045,
42
+ 0.0
43
+ ],
44
+ "max": [
45
+ 0.6283698678016663,
46
+ 0.4420371949672699,
47
+ 1.051027774810791,
48
+ 0.9999999403953552,
49
+ 0.9999995231628418,
50
+ 1.0,
51
+ 0.9999995231628418,
52
+ 1.0,
53
+ 0.9999996423721313,
54
+ 0.03999999910593033
55
+ ],
56
+ "mean": [
57
+ 0.4040228037735655,
58
+ -0.05247618696989652,
59
+ 0.6993459771218912,
60
+ 0.07011111991957117,
61
+ -0.03520558904460197,
62
+ 0.14470242846953205,
63
+ 0.05491166606739672,
64
+ -0.19073058893378805,
65
+ -0.06988117661871228,
66
+ 0.021167300266524632
67
+ ],
68
+ "std": [
69
+ 0.11861927913357073,
70
+ 0.11318943557281885,
71
+ 0.17849160646987516,
72
+ 0.4812528791886511,
73
+ 0.7487196258736938,
74
+ 0.4251126804030219,
75
+ 0.47797229807769137,
76
+ 0.6267695346407992,
77
+ 0.5782954306815554,
78
+ 0.01996590431050817
79
+ ],
80
+ "count": [
81
+ 218367
82
+ ],
83
+ "q01": [
84
+ 0.23988881827972064,
85
+ -0.10387461883034811,
86
+ 0.6297243923476871,
87
+ -0.43622841955353464,
88
+ -0.6885486527617446,
89
+ -0.14103555388039019,
90
+ -0.4327210063987595,
91
+ -0.9304790915716145,
92
+ -0.3465462258490417,
93
+ -1.000000013351432e-10
94
+ ],
95
+ "q10": [
96
+ 0.3042306436755216,
97
+ -0.09609991620622624,
98
+ 0.6394063072781145,
99
+ -0.3351306352998669,
100
+ -0.5751963015002924,
101
+ -0.08490621283307866,
102
+ -0.33849701676224275,
103
+ -0.8340557712529717,
104
+ -0.29760290790884514,
105
+ -1.000000013351432e-10
106
+ ],
107
+ "q50": [
108
+ 0.4160225263528291,
109
+ -0.06267459152989589,
110
+ 0.6811777491060801,
111
+ 0.0543536902230377,
112
+ -0.14012251058819217,
113
+ 0.04772394280004727,
114
+ 0.03838119138716219,
115
+ -0.14865169094195635,
116
+ -0.2073069401172425,
117
+ 0.020772594934870346
118
+ ],
119
+ "q90": [
120
+ 0.46641008755679864,
121
+ 0.020233143101890064,
122
+ 0.8095135405259374,
123
+ 0.48172401592474906,
124
+ 0.7398737043070319,
125
+ 0.5970841215636613,
126
+ 0.4674752875999555,
127
+ 0.36281754777672676,
128
+ 0.5033361883834496,
129
+ 0.039998362213352394
130
+ ],
131
+ "q99": [
132
+ 0.47320723171849244,
133
+ 0.04114262476135705,
134
+ 0.8467967259270389,
135
+ 0.5867781409980988,
136
+ 0.8529584859257228,
137
+ 0.7953936054341789,
138
+ 0.5768269858732034,
139
+ 0.49353349323389695,
140
+ 0.7168914440776504,
141
+ 0.03999983541667277
142
+ ]
143
+ }
144
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/phase_schedule.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "step_interval": "[start_step, until_step)",
4
+ "total_steps": 60000,
5
+ "optimizer_lifecycle": "single_continuous_optimizer",
6
+ "scheduler_lifecycle": "single_continuous_scheduler",
7
+ "phases": [
8
+ {
9
+ "index": 0,
10
+ "name": "rscl_rkd_cosine_decay_60K",
11
+ "start_step": 0,
12
+ "until_step": 60000,
13
+ "policy": {
14
+ "rkd_enabled": true,
15
+ "rkd_fm_loss_weight": 1.0,
16
+ "rkd_loss_weight": 1.0,
17
+ "rkd_loss_weight_end": 0.0,
18
+ "rkd_loss_weight_schedule": "cosine",
19
+ "rkd_distance_weight": 1.0,
20
+ "rkd_angle_weight": 2.0
21
+ },
22
+ "trainable": {
23
+ "tune_llm": false,
24
+ "tune_visual": false,
25
+ "tune_top_llm_layers": 0,
26
+ "tune_projector": true,
27
+ "tune_diffusion_model": true,
28
+ "tune_vlln": true
29
+ }
30
+ }
31
+ ]
32
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/resolved_config.yaml ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dataset:
2
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
3
+ root: /home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2
4
+ policy:
5
+ type: groot_rkd
6
+ device: cuda
7
+ chunk_size: 16
8
+ n_action_steps: 16
9
+ push_to_hub: false
10
+ warmup_steps: 3000
11
+ tune_llm: false
12
+ tune_visual: false
13
+ tune_top_llm_layers: 0
14
+ action_expert_num_layers: 16
15
+ tune_projector: true
16
+ tune_diffusion_model: true
17
+ tune_vlln: true
18
+ rkd_enabled: true
19
+ rkd_vae_checkpoint: /home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
20
+ rkd_student_pooling: attention
21
+ rkd_student_attention_hidden_dim: 512
22
+ rkd_student_attention_output_dim: 512
23
+ rkd_fm_loss_weight: 1.0
24
+ rkd_loss_weight: 1.0
25
+ rkd_distance_weight: 1.0
26
+ rkd_angle_weight: 2.0
27
+ seed: 42
28
+ batch_size: 64
29
+ steps: 60000
30
+ log_freq: 200
31
+ output_dir: /home/ext_minje/groot_insight/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
32
+ job_name: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
33
+ wandb:
34
+ enable: true
35
+ disable_artifact: true
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/source_config.yaml ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # INSIGHT GR00T N1.7 RKD training with RS-CL-style full-training regularization.
2
+ # Common RKD values live in the top-level policy; the phase contains schedule overrides only.
3
+ insight:
4
+ checkpoint_interval: 5K
5
+ checkpoint_slots: 2
6
+ # Options: joint_abs | ee_abs_quat | ee_abs_rot6d | ee_delta
7
+ action_space: ee_abs_rot6d
8
+ # Options: coarse | detailed
9
+ prompt_set: detailed
10
+ save_steps: ["10K", "30K", "60K"]
11
+ phases:
12
+ - name: rscl_rkd_cosine_decay_60K
13
+ until_step: 60K
14
+ policy:
15
+ # Inherits RKD=1.0 below and follows a half-cosine decay to zero.
16
+ rkd_loss_weight_end: 0.0
17
+ rkd_loss_weight_schedule: cosine
18
+
19
+ dataset:
20
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
21
+ root: ${HOME}/INSIGHTfixposV4_filtered_multispace_v2
22
+
23
+ policy:
24
+ type: groot_rkd
25
+ device: cuda
26
+ chunk_size: 16
27
+ n_action_steps: 16
28
+ push_to_hub: false
29
+ # RS-CL GR00T uses exactly 3K warmup updates.
30
+ warmup_steps: 3000
31
+
32
+ # VLM
33
+ tune_llm: false
34
+ tune_visual: false
35
+ tune_top_llm_layers: 0
36
+ # Action Expert
37
+ action_expert_num_layers: 16
38
+ tune_projector: true
39
+ tune_diffusion_model: true
40
+ tune_vlln: true
41
+
42
+ # Frozen TimewarpVAE teacher and common RKD objective.
43
+ rkd_enabled: true
44
+ rkd_vae_checkpoint: ${HOME}/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
45
+ rkd_student_pooling: attention # attention | mean
46
+ rkd_student_attention_hidden_dim: 512
47
+ rkd_student_attention_output_dim: 512
48
+ rkd_fm_loss_weight: 1.0
49
+ rkd_loss_weight: 1.0
50
+ rkd_distance_weight: 1.0
51
+ rkd_angle_weight: 2.0
52
+
53
+ seed: 42
54
+ # Single GPU/process, so this is also the global batch size.
55
+ batch_size: 64
56
+ steps: 60000
57
+ log_freq: 200
58
+ output_dir: ${HOME}/groot_insight/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
59
+ job_name: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
60
+
61
+ wandb:
62
+ enable: true
63
+ disable_artifact: true
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,361 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 1.8865999845374793e-06,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 1e-05,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "initial_lr": 0.0001,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68,
88
+ 69,
89
+ 70,
90
+ 71,
91
+ 72,
92
+ 73,
93
+ 74,
94
+ 75,
95
+ 76,
96
+ 77,
97
+ 78,
98
+ 79,
99
+ 80,
100
+ 81,
101
+ 82,
102
+ 83,
103
+ 84,
104
+ 85,
105
+ 86,
106
+ 87,
107
+ 88,
108
+ 89,
109
+ 90,
110
+ 91,
111
+ 92,
112
+ 93,
113
+ 94,
114
+ 95,
115
+ 96,
116
+ 97,
117
+ 98,
118
+ 99,
119
+ 100,
120
+ 101,
121
+ 102,
122
+ 103,
123
+ 104,
124
+ 105,
125
+ 106,
126
+ 107,
127
+ 108,
128
+ 109,
129
+ 110,
130
+ 111,
131
+ 112,
132
+ 113,
133
+ 114,
134
+ 115,
135
+ 116,
136
+ 117,
137
+ 118,
138
+ 119,
139
+ 120,
140
+ 121,
141
+ 122,
142
+ 123,
143
+ 124,
144
+ 125,
145
+ 126,
146
+ 127,
147
+ 128,
148
+ 129,
149
+ 130,
150
+ 131,
151
+ 132,
152
+ 133,
153
+ 134,
154
+ 135,
155
+ 136,
156
+ 137,
157
+ 138,
158
+ 139,
159
+ 140,
160
+ 141,
161
+ 142,
162
+ 143,
163
+ 144,
164
+ 145,
165
+ 146,
166
+ 147,
167
+ 148,
168
+ 149,
169
+ 150,
170
+ 151,
171
+ 152,
172
+ 153,
173
+ 154,
174
+ 155,
175
+ 156,
176
+ 157,
177
+ 158
178
+ ]
179
+ },
180
+ {
181
+ "weight_decay": 0.0,
182
+ "lr": 1.8865999845374793e-06,
183
+ "betas": [
184
+ 0.9,
185
+ 0.999
186
+ ],
187
+ "eps": 1e-08,
188
+ "amsgrad": false,
189
+ "maximize": false,
190
+ "foreach": null,
191
+ "capturable": false,
192
+ "differentiable": false,
193
+ "fused": null,
194
+ "decoupled_weight_decay": true,
195
+ "initial_lr": 0.0001,
196
+ "params": [
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267,
306
+ 268,
307
+ 269,
308
+ 270,
309
+ 271,
310
+ 272,
311
+ 273,
312
+ 274,
313
+ 275,
314
+ 276,
315
+ 277,
316
+ 278,
317
+ 279,
318
+ 280,
319
+ 281,
320
+ 282,
321
+ 283,
322
+ 284,
323
+ 285,
324
+ 286,
325
+ 287,
326
+ 288,
327
+ 289,
328
+ 290,
329
+ 291,
330
+ 292,
331
+ 293,
332
+ 294,
333
+ 295,
334
+ 296,
335
+ 297,
336
+ 298,
337
+ 299,
338
+ 300,
339
+ 301,
340
+ 302,
341
+ 303,
342
+ 304,
343
+ 305,
344
+ 306,
345
+ 307,
346
+ 308,
347
+ 309,
348
+ 310,
349
+ 311,
350
+ 312,
351
+ 313,
352
+ 314,
353
+ 315,
354
+ 316,
355
+ 317,
356
+ 318,
357
+ 319,
358
+ 320
359
+ ]
360
+ }
361
+ ]
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00/training_state/training_step.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "step": 55000,
3
+ "num_processes": 1,
4
+ "batch_size": 64
5
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/action_space_manifest.json ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 2,
3
+ "action_space": "ee_abs_rot6d",
4
+ "action_views": [
5
+ "ee_abs_rot6d"
6
+ ],
7
+ "source_feature": "action.ee_abs_rot6d",
8
+ "canonical_feature": "action",
9
+ "dtype": "float32",
10
+ "shape": [
11
+ 10
12
+ ],
13
+ "names": [
14
+ "x",
15
+ "y",
16
+ "z",
17
+ "r_col0_x",
18
+ "r_col0_y",
19
+ "r_col0_z",
20
+ "r_col1_x",
21
+ "r_col1_y",
22
+ "r_col1_z",
23
+ "gripper"
24
+ ],
25
+ "dataset_repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
26
+ "dataset_revision": "v3.0",
27
+ "dataset_total_frames": 218367,
28
+ "dataset_total_episodes": 4930,
29
+ "dataset_fps": 10,
30
+ "stats_sha256": "71698c7337070aa8267df5aea7ec08bc00d3925510f2649bc20c9201eeb669e4",
31
+ "stats": {
32
+ "min": [
33
+ -0.21784618496894836,
34
+ -0.4474811851978302,
35
+ 0.1702737659215927,
36
+ -0.9999986290931702,
37
+ -0.9999991655349731,
38
+ -0.9999999403953552,
39
+ -0.9999990463256836,
40
+ -0.9999998211860657,
41
+ -0.9999988079071045,
42
+ 0.0
43
+ ],
44
+ "max": [
45
+ 0.6283698678016663,
46
+ 0.4420371949672699,
47
+ 1.051027774810791,
48
+ 0.9999999403953552,
49
+ 0.9999995231628418,
50
+ 1.0,
51
+ 0.9999995231628418,
52
+ 1.0,
53
+ 0.9999996423721313,
54
+ 0.03999999910593033
55
+ ],
56
+ "mean": [
57
+ 0.4040228037735655,
58
+ -0.05247618696989652,
59
+ 0.6993459771218912,
60
+ 0.07011111991957117,
61
+ -0.03520558904460197,
62
+ 0.14470242846953205,
63
+ 0.05491166606739672,
64
+ -0.19073058893378805,
65
+ -0.06988117661871228,
66
+ 0.021167300266524632
67
+ ],
68
+ "std": [
69
+ 0.11861927913357073,
70
+ 0.11318943557281885,
71
+ 0.17849160646987516,
72
+ 0.4812528791886511,
73
+ 0.7487196258736938,
74
+ 0.4251126804030219,
75
+ 0.47797229807769137,
76
+ 0.6267695346407992,
77
+ 0.5782954306815554,
78
+ 0.01996590431050817
79
+ ],
80
+ "count": [
81
+ 218367
82
+ ],
83
+ "q01": [
84
+ 0.23988881827972064,
85
+ -0.10387461883034811,
86
+ 0.6297243923476871,
87
+ -0.43622841955353464,
88
+ -0.6885486527617446,
89
+ -0.14103555388039019,
90
+ -0.4327210063987595,
91
+ -0.9304790915716145,
92
+ -0.3465462258490417,
93
+ -1.000000013351432e-10
94
+ ],
95
+ "q10": [
96
+ 0.3042306436755216,
97
+ -0.09609991620622624,
98
+ 0.6394063072781145,
99
+ -0.3351306352998669,
100
+ -0.5751963015002924,
101
+ -0.08490621283307866,
102
+ -0.33849701676224275,
103
+ -0.8340557712529717,
104
+ -0.29760290790884514,
105
+ -1.000000013351432e-10
106
+ ],
107
+ "q50": [
108
+ 0.4160225263528291,
109
+ -0.06267459152989589,
110
+ 0.6811777491060801,
111
+ 0.0543536902230377,
112
+ -0.14012251058819217,
113
+ 0.04772394280004727,
114
+ 0.03838119138716219,
115
+ -0.14865169094195635,
116
+ -0.2073069401172425,
117
+ 0.020772594934870346
118
+ ],
119
+ "q90": [
120
+ 0.46641008755679864,
121
+ 0.020233143101890064,
122
+ 0.8095135405259374,
123
+ 0.48172401592474906,
124
+ 0.7398737043070319,
125
+ 0.5970841215636613,
126
+ 0.4674752875999555,
127
+ 0.36281754777672676,
128
+ 0.5033361883834496,
129
+ 0.039998362213352394
130
+ ],
131
+ "q99": [
132
+ 0.47320723171849244,
133
+ 0.04114262476135705,
134
+ 0.8467967259270389,
135
+ 0.5867781409980988,
136
+ 0.8529584859257228,
137
+ 0.7953936054341789,
138
+ 0.5768269858732034,
139
+ 0.49353349323389695,
140
+ 0.7168914440776504,
141
+ 0.03999983541667277
142
+ ]
143
+ }
144
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/phase_schedule.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "step_interval": "[start_step, until_step)",
4
+ "total_steps": 60000,
5
+ "optimizer_lifecycle": "single_continuous_optimizer",
6
+ "scheduler_lifecycle": "single_continuous_scheduler",
7
+ "phases": [
8
+ {
9
+ "index": 0,
10
+ "name": "rscl_rkd_cosine_decay_60K",
11
+ "start_step": 0,
12
+ "until_step": 60000,
13
+ "policy": {
14
+ "rkd_enabled": true,
15
+ "rkd_fm_loss_weight": 1.0,
16
+ "rkd_loss_weight": 1.0,
17
+ "rkd_loss_weight_end": 0.0,
18
+ "rkd_loss_weight_schedule": "cosine",
19
+ "rkd_distance_weight": 1.0,
20
+ "rkd_angle_weight": 2.0
21
+ },
22
+ "trainable": {
23
+ "tune_llm": false,
24
+ "tune_visual": false,
25
+ "tune_top_llm_layers": 0,
26
+ "tune_projector": true,
27
+ "tune_diffusion_model": true,
28
+ "tune_vlln": true
29
+ }
30
+ }
31
+ ]
32
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/action_space_manifest.json ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 2,
3
+ "action_space": "ee_abs_rot6d",
4
+ "action_views": [
5
+ "ee_abs_rot6d"
6
+ ],
7
+ "source_feature": "action.ee_abs_rot6d",
8
+ "canonical_feature": "action",
9
+ "dtype": "float32",
10
+ "shape": [
11
+ 10
12
+ ],
13
+ "names": [
14
+ "x",
15
+ "y",
16
+ "z",
17
+ "r_col0_x",
18
+ "r_col0_y",
19
+ "r_col0_z",
20
+ "r_col1_x",
21
+ "r_col1_y",
22
+ "r_col1_z",
23
+ "gripper"
24
+ ],
25
+ "dataset_repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
26
+ "dataset_revision": "v3.0",
27
+ "dataset_total_frames": 218367,
28
+ "dataset_total_episodes": 4930,
29
+ "dataset_fps": 10,
30
+ "stats_sha256": "71698c7337070aa8267df5aea7ec08bc00d3925510f2649bc20c9201eeb669e4",
31
+ "stats": {
32
+ "min": [
33
+ -0.21784618496894836,
34
+ -0.4474811851978302,
35
+ 0.1702737659215927,
36
+ -0.9999986290931702,
37
+ -0.9999991655349731,
38
+ -0.9999999403953552,
39
+ -0.9999990463256836,
40
+ -0.9999998211860657,
41
+ -0.9999988079071045,
42
+ 0.0
43
+ ],
44
+ "max": [
45
+ 0.6283698678016663,
46
+ 0.4420371949672699,
47
+ 1.051027774810791,
48
+ 0.9999999403953552,
49
+ 0.9999995231628418,
50
+ 1.0,
51
+ 0.9999995231628418,
52
+ 1.0,
53
+ 0.9999996423721313,
54
+ 0.03999999910593033
55
+ ],
56
+ "mean": [
57
+ 0.4040228037735655,
58
+ -0.05247618696989652,
59
+ 0.6993459771218912,
60
+ 0.07011111991957117,
61
+ -0.03520558904460197,
62
+ 0.14470242846953205,
63
+ 0.05491166606739672,
64
+ -0.19073058893378805,
65
+ -0.06988117661871228,
66
+ 0.021167300266524632
67
+ ],
68
+ "std": [
69
+ 0.11861927913357073,
70
+ 0.11318943557281885,
71
+ 0.17849160646987516,
72
+ 0.4812528791886511,
73
+ 0.7487196258736938,
74
+ 0.4251126804030219,
75
+ 0.47797229807769137,
76
+ 0.6267695346407992,
77
+ 0.5782954306815554,
78
+ 0.01996590431050817
79
+ ],
80
+ "count": [
81
+ 218367
82
+ ],
83
+ "q01": [
84
+ 0.23988881827972064,
85
+ -0.10387461883034811,
86
+ 0.6297243923476871,
87
+ -0.43622841955353464,
88
+ -0.6885486527617446,
89
+ -0.14103555388039019,
90
+ -0.4327210063987595,
91
+ -0.9304790915716145,
92
+ -0.3465462258490417,
93
+ -1.000000013351432e-10
94
+ ],
95
+ "q10": [
96
+ 0.3042306436755216,
97
+ -0.09609991620622624,
98
+ 0.6394063072781145,
99
+ -0.3351306352998669,
100
+ -0.5751963015002924,
101
+ -0.08490621283307866,
102
+ -0.33849701676224275,
103
+ -0.8340557712529717,
104
+ -0.29760290790884514,
105
+ -1.000000013351432e-10
106
+ ],
107
+ "q50": [
108
+ 0.4160225263528291,
109
+ -0.06267459152989589,
110
+ 0.6811777491060801,
111
+ 0.0543536902230377,
112
+ -0.14012251058819217,
113
+ 0.04772394280004727,
114
+ 0.03838119138716219,
115
+ -0.14865169094195635,
116
+ -0.2073069401172425,
117
+ 0.020772594934870346
118
+ ],
119
+ "q90": [
120
+ 0.46641008755679864,
121
+ 0.020233143101890064,
122
+ 0.8095135405259374,
123
+ 0.48172401592474906,
124
+ 0.7398737043070319,
125
+ 0.5970841215636613,
126
+ 0.4674752875999555,
127
+ 0.36281754777672676,
128
+ 0.5033361883834496,
129
+ 0.039998362213352394
130
+ ],
131
+ "q99": [
132
+ 0.47320723171849244,
133
+ 0.04114262476135705,
134
+ 0.8467967259270389,
135
+ 0.5867781409980988,
136
+ 0.8529584859257228,
137
+ 0.7953936054341789,
138
+ 0.5768269858732034,
139
+ 0.49353349323389695,
140
+ 0.7168914440776504,
141
+ 0.03999983541667277
142
+ ]
143
+ }
144
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/config.json ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "groot_rkd",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 16
9
+ ]
10
+ },
11
+ "observation.images.wrist": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 224,
16
+ 224
17
+ ]
18
+ },
19
+ "observation.images.right_shoulder": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 224,
24
+ 224
25
+ ]
26
+ },
27
+ "observation.images.guide": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 224,
32
+ 224
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 10
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": false,
46
+ "use_peft": false,
47
+ "push_to_hub": false,
48
+ "repo_id": null,
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model",
53
+ "pretrained_revision": null,
54
+ "chunk_size": 16,
55
+ "n_action_steps": 16,
56
+ "max_state_dim": 132,
57
+ "max_action_dim": 132,
58
+ "normalization_mapping": {
59
+ "VISUAL": "IDENTITY",
60
+ "STATE": "IDENTITY",
61
+ "ACTION": "IDENTITY"
62
+ },
63
+ "base_model_path": "nvidia/GR00T-N1.7-3B",
64
+ "action_decode_transform": null,
65
+ "embodiment_tag": "new_embodiment",
66
+ "tune_llm": false,
67
+ "tune_visual": false,
68
+ "tune_projector": true,
69
+ "tune_diffusion_model": true,
70
+ "tune_vlln": true,
71
+ "tune_top_llm_layers": 0,
72
+ "num_inference_timesteps": null,
73
+ "rtc_ramp_rate": null,
74
+ "use_flash_attention": false,
75
+ "use_relative_actions": false,
76
+ "relative_exclude_joints": [],
77
+ "optimizer_lr": 0.0001,
78
+ "optimizer_betas": [
79
+ 0.9,
80
+ 0.999
81
+ ],
82
+ "optimizer_eps": 1e-08,
83
+ "optimizer_weight_decay": 1e-05,
84
+ "warmup_steps": 3000,
85
+ "warmup_ratio": 0.05,
86
+ "use_bf16": true,
87
+ "model_params_fp32": true,
88
+ "image_size": [
89
+ 256,
90
+ 256
91
+ ],
92
+ "tokenizer_assets_repo": null,
93
+ "lora_rank": 0,
94
+ "lora_alpha": 16,
95
+ "lora_dropout": 0.1,
96
+ "lora_full_model": false,
97
+ "video_backend": "decord",
98
+ "balance_dataset_weights": true,
99
+ "balance_trajectory_weights": true,
100
+ "dataset_paths": null,
101
+ "output_dir": "./tmp/gr00t",
102
+ "save_steps": 1000,
103
+ "max_steps": 10000,
104
+ "batch_size": 32,
105
+ "dataloader_num_workers": 8,
106
+ "report_to": "wandb",
107
+ "resume": false,
108
+ "action_expert_num_layers": 16,
109
+ "rkd_enabled": true,
110
+ "rkd_vae_checkpoint": "/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11",
111
+ "rkd_student_pooling": "attention",
112
+ "rkd_student_attention_hidden_dim": 512,
113
+ "rkd_student_attention_output_dim": 512,
114
+ "rkd_fm_loss_weight": 1.0,
115
+ "rkd_loss_weight": 1.0,
116
+ "rkd_distance_weight": 1.0,
117
+ "rkd_angle_weight": 2.0
118
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/phase_schedule.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "step_interval": "[start_step, until_step)",
4
+ "total_steps": 60000,
5
+ "optimizer_lifecycle": "single_continuous_optimizer",
6
+ "scheduler_lifecycle": "single_continuous_scheduler",
7
+ "phases": [
8
+ {
9
+ "index": 0,
10
+ "name": "rscl_rkd_cosine_decay_60K",
11
+ "start_step": 0,
12
+ "until_step": 60000,
13
+ "policy": {
14
+ "rkd_enabled": true,
15
+ "rkd_fm_loss_weight": 1.0,
16
+ "rkd_loss_weight": 1.0,
17
+ "rkd_loss_weight_end": 0.0,
18
+ "rkd_loss_weight_schedule": "cosine",
19
+ "rkd_distance_weight": 1.0,
20
+ "rkd_angle_weight": 2.0
21
+ },
22
+ "trainable": {
23
+ "tune_llm": false,
24
+ "tune_visual": false,
25
+ "tune_top_llm_layers": 0,
26
+ "tune_projector": true,
27
+ "tune_diffusion_model": true,
28
+ "tune_vlln": true
29
+ }
30
+ }
31
+ ]
32
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/prompt_manifest.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "prompt_set": "detailed",
4
+ "source_file": "task_prompts.json",
5
+ "source_sha256": "0ba48fb372da9f49dcfcd7dcc2da8cce6ce991d3661060e664a104778d1ffacc",
6
+ "mapping_sha256": "f71954786e94cdeceb135dde09fdc6b65db0b87ab3b09d25060129d0f0fde018",
7
+ "dataset_repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
8
+ "dataset_revision": "v3.0",
9
+ "task_count": 13,
10
+ "task_codes": [
11
+ "5f",
12
+ "5g",
13
+ "1ext",
14
+ "3b",
15
+ "5b",
16
+ "3a",
17
+ "3c",
18
+ "5h",
19
+ "5d",
20
+ "3d",
21
+ "5e",
22
+ "5a",
23
+ "5c"
24
+ ],
25
+ "prompts": {
26
+ "5f": "Close the bottle in clockwise direction.",
27
+ "5g": "Grip the cap on the sides indicated by the 'squeeze' arrow and open the bottle in clockwise direction.",
28
+ "1ext": "Find the arrow guide and open the indicated drawer.",
29
+ "3b": "Open the door, rotate counter-clockwise and push.",
30
+ "5b": "Open the bottle in counter-clockwise direction.",
31
+ "3a": "Open the door, rotate clockwise and push.",
32
+ "3c": "Open the door, rotate clockwise and pull.",
33
+ "5h": "Close the bottle in counter-clockwise direction.",
34
+ "5d": "Close the bottle in clockwise direction.",
35
+ "3d": "Open the door, rotate counter-clockwise and pull.",
36
+ "5e": "Close the bottle in counter-clockwise direction.",
37
+ "5a": "Grip the cap on the sides indicated by the 'squeeze' arrow and open the bottle in counter-clockwise direction.",
38
+ "5c": "Open the bottle in clockwise direction."
39
+ }
40
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/train_config.json ADDED
@@ -0,0 +1,269 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
4
+ "root": "/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "return_uint8": false,
81
+ "depth_output_unit": "mm",
82
+ "streaming": false,
83
+ "eval_split": 0.0
84
+ },
85
+ "env": null,
86
+ "policy": {
87
+ "type": "groot_rkd",
88
+ "n_obs_steps": 1,
89
+ "input_features": {
90
+ "observation.state": {
91
+ "type": "STATE",
92
+ "shape": [
93
+ 16
94
+ ]
95
+ },
96
+ "observation.images.wrist": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 3,
100
+ 224,
101
+ 224
102
+ ]
103
+ },
104
+ "observation.images.right_shoulder": {
105
+ "type": "VISUAL",
106
+ "shape": [
107
+ 3,
108
+ 224,
109
+ 224
110
+ ]
111
+ },
112
+ "observation.images.guide": {
113
+ "type": "VISUAL",
114
+ "shape": [
115
+ 3,
116
+ 224,
117
+ 224
118
+ ]
119
+ }
120
+ },
121
+ "output_features": {
122
+ "action": {
123
+ "type": "ACTION",
124
+ "shape": [
125
+ 10
126
+ ]
127
+ }
128
+ },
129
+ "device": "cuda",
130
+ "use_amp": false,
131
+ "use_peft": false,
132
+ "push_to_hub": false,
133
+ "repo_id": null,
134
+ "private": null,
135
+ "tags": null,
136
+ "license": null,
137
+ "pretrained_path": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model",
138
+ "pretrained_revision": null,
139
+ "chunk_size": 16,
140
+ "n_action_steps": 16,
141
+ "max_state_dim": 132,
142
+ "max_action_dim": 132,
143
+ "normalization_mapping": {
144
+ "VISUAL": "IDENTITY",
145
+ "STATE": "IDENTITY",
146
+ "ACTION": "IDENTITY"
147
+ },
148
+ "base_model_path": "nvidia/GR00T-N1.7-3B",
149
+ "action_decode_transform": null,
150
+ "embodiment_tag": "new_embodiment",
151
+ "tune_llm": false,
152
+ "tune_visual": false,
153
+ "tune_projector": true,
154
+ "tune_diffusion_model": true,
155
+ "tune_vlln": true,
156
+ "tune_top_llm_layers": 0,
157
+ "num_inference_timesteps": null,
158
+ "rtc_ramp_rate": null,
159
+ "use_flash_attention": false,
160
+ "use_relative_actions": false,
161
+ "relative_exclude_joints": [],
162
+ "optimizer_lr": 0.0001,
163
+ "optimizer_betas": [
164
+ 0.9,
165
+ 0.999
166
+ ],
167
+ "optimizer_eps": 1e-08,
168
+ "optimizer_weight_decay": 1e-05,
169
+ "warmup_steps": 3000,
170
+ "warmup_ratio": 0.05,
171
+ "use_bf16": true,
172
+ "model_params_fp32": true,
173
+ "image_size": [
174
+ 256,
175
+ 256
176
+ ],
177
+ "tokenizer_assets_repo": null,
178
+ "lora_rank": 0,
179
+ "lora_alpha": 16,
180
+ "lora_dropout": 0.1,
181
+ "lora_full_model": false,
182
+ "video_backend": "decord",
183
+ "balance_dataset_weights": true,
184
+ "balance_trajectory_weights": true,
185
+ "dataset_paths": null,
186
+ "output_dir": "./tmp/gr00t",
187
+ "save_steps": 1000,
188
+ "max_steps": 10000,
189
+ "batch_size": 32,
190
+ "dataloader_num_workers": 8,
191
+ "report_to": "wandb",
192
+ "resume": false,
193
+ "action_expert_num_layers": 16,
194
+ "rkd_enabled": true,
195
+ "rkd_vae_checkpoint": "/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11",
196
+ "rkd_student_pooling": "attention",
197
+ "rkd_student_attention_hidden_dim": 512,
198
+ "rkd_student_attention_output_dim": 512,
199
+ "rkd_fm_loss_weight": 1.0,
200
+ "rkd_loss_weight": 1.0,
201
+ "rkd_distance_weight": 1.0,
202
+ "rkd_angle_weight": 2.0
203
+ },
204
+ "reward_model": null,
205
+ "output_dir": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K",
206
+ "job_name": "INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60",
207
+ "resume": true,
208
+ "seed": 42,
209
+ "cudnn_deterministic": false,
210
+ "num_workers": 4,
211
+ "batch_size": 64,
212
+ "prefetch_factor": 4,
213
+ "persistent_workers": true,
214
+ "steps": 60000,
215
+ "env_eval_freq": 20000,
216
+ "log_freq": 200,
217
+ "eval_steps": 0,
218
+ "max_eval_samples": 0,
219
+ "tolerance_s": 0.0001,
220
+ "save_checkpoint": true,
221
+ "save_freq": 20000,
222
+ "use_policy_training_preset": true,
223
+ "optimizer": {
224
+ "type": "adamw",
225
+ "lr": 0.0001,
226
+ "weight_decay": 1e-05,
227
+ "grad_clip_norm": 1.0,
228
+ "betas": [
229
+ 0.9,
230
+ 0.999
231
+ ],
232
+ "eps": 1e-08
233
+ },
234
+ "scheduler": {
235
+ "type": "diffuser",
236
+ "num_warmup_steps": 3000,
237
+ "name": "cosine"
238
+ },
239
+ "eval": {
240
+ "n_episodes": 50,
241
+ "batch_size": 50,
242
+ "use_async_envs": true,
243
+ "recording": false,
244
+ "recording_repo_id": null,
245
+ "recording_private": false
246
+ },
247
+ "wandb": {
248
+ "enable": true,
249
+ "disable_artifact": true,
250
+ "project": "lerobot",
251
+ "entity": null,
252
+ "notes": null,
253
+ "run_id": "rs1egwv8",
254
+ "mode": null,
255
+ "add_tags": true
256
+ },
257
+ "peft": null,
258
+ "job": {
259
+ "target": null,
260
+ "image": "huggingface/lerobot-gpu:latest",
261
+ "timeout": "2d",
262
+ "detach": false,
263
+ "tags": []
264
+ },
265
+ "save_checkpoint_to_hub": false,
266
+ "sample_weighting": null,
267
+ "rename_map": {},
268
+ "checkpoint_path": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01"
269
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/prompt_manifest.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "prompt_set": "detailed",
4
+ "source_file": "task_prompts.json",
5
+ "source_sha256": "0ba48fb372da9f49dcfcd7dcc2da8cce6ce991d3661060e664a104778d1ffacc",
6
+ "mapping_sha256": "f71954786e94cdeceb135dde09fdc6b65db0b87ab3b09d25060129d0f0fde018",
7
+ "dataset_repo_id": "Whalswp/INSIGHTfixposV4_filtered_multispace_v2",
8
+ "dataset_revision": "v3.0",
9
+ "task_count": 13,
10
+ "task_codes": [
11
+ "5f",
12
+ "5g",
13
+ "1ext",
14
+ "3b",
15
+ "5b",
16
+ "3a",
17
+ "3c",
18
+ "5h",
19
+ "5d",
20
+ "3d",
21
+ "5e",
22
+ "5a",
23
+ "5c"
24
+ ],
25
+ "prompts": {
26
+ "5f": "Close the bottle in clockwise direction.",
27
+ "5g": "Grip the cap on the sides indicated by the 'squeeze' arrow and open the bottle in clockwise direction.",
28
+ "1ext": "Find the arrow guide and open the indicated drawer.",
29
+ "3b": "Open the door, rotate counter-clockwise and push.",
30
+ "5b": "Open the bottle in counter-clockwise direction.",
31
+ "3a": "Open the door, rotate clockwise and push.",
32
+ "3c": "Open the door, rotate clockwise and pull.",
33
+ "5h": "Close the bottle in counter-clockwise direction.",
34
+ "5d": "Close the bottle in clockwise direction.",
35
+ "3d": "Open the door, rotate counter-clockwise and pull.",
36
+ "5e": "Close the bottle in counter-clockwise direction.",
37
+ "5a": "Grip the cap on the sides indicated by the 'squeeze' arrow and open the bottle in counter-clockwise direction.",
38
+ "5c": "Open the bottle in clockwise direction."
39
+ }
40
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/resolved_config.yaml ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dataset:
2
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
3
+ root: /home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2
4
+ policy:
5
+ type: groot_rkd
6
+ device: cuda
7
+ chunk_size: 16
8
+ n_action_steps: 16
9
+ push_to_hub: false
10
+ warmup_steps: 3000
11
+ tune_llm: false
12
+ tune_visual: false
13
+ tune_top_llm_layers: 0
14
+ action_expert_num_layers: 16
15
+ tune_projector: true
16
+ tune_diffusion_model: true
17
+ tune_vlln: true
18
+ rkd_enabled: true
19
+ rkd_vae_checkpoint: /home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
20
+ rkd_student_pooling: attention
21
+ rkd_student_attention_hidden_dim: 512
22
+ rkd_student_attention_output_dim: 512
23
+ rkd_fm_loss_weight: 1.0
24
+ rkd_loss_weight: 1.0
25
+ rkd_distance_weight: 1.0
26
+ rkd_angle_weight: 2.0
27
+ seed: 42
28
+ batch_size: 64
29
+ steps: 60000
30
+ log_freq: 200
31
+ output_dir: /home/ext_minje/groot_insight/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
32
+ job_name: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
33
+ wandb:
34
+ enable: true
35
+ disable_artifact: true
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/source_config.yaml ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # INSIGHT GR00T N1.7 RKD training with RS-CL-style full-training regularization.
2
+ # Common RKD values live in the top-level policy; the phase contains schedule overrides only.
3
+ insight:
4
+ checkpoint_interval: 5K
5
+ checkpoint_slots: 2
6
+ # Options: joint_abs | ee_abs_quat | ee_abs_rot6d | ee_delta
7
+ action_space: ee_abs_rot6d
8
+ # Options: coarse | detailed
9
+ prompt_set: detailed
10
+ save_steps: ["10K", "30K", "60K"]
11
+ phases:
12
+ - name: rscl_rkd_cosine_decay_60K
13
+ until_step: 60K
14
+ policy:
15
+ # Inherits RKD=1.0 below and follows a half-cosine decay to zero.
16
+ rkd_loss_weight_end: 0.0
17
+ rkd_loss_weight_schedule: cosine
18
+
19
+ dataset:
20
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
21
+ root: ${HOME}/INSIGHTfixposV4_filtered_multispace_v2
22
+
23
+ policy:
24
+ type: groot_rkd
25
+ device: cuda
26
+ chunk_size: 16
27
+ n_action_steps: 16
28
+ push_to_hub: false
29
+ # RS-CL GR00T uses exactly 3K warmup updates.
30
+ warmup_steps: 3000
31
+
32
+ # VLM
33
+ tune_llm: false
34
+ tune_visual: false
35
+ tune_top_llm_layers: 0
36
+ # Action Expert
37
+ action_expert_num_layers: 16
38
+ tune_projector: true
39
+ tune_diffusion_model: true
40
+ tune_vlln: true
41
+
42
+ # Frozen TimewarpVAE teacher and common RKD objective.
43
+ rkd_enabled: true
44
+ rkd_vae_checkpoint: ${HOME}/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
45
+ rkd_student_pooling: attention # attention | mean
46
+ rkd_student_attention_hidden_dim: 512
47
+ rkd_student_attention_output_dim: 512
48
+ rkd_fm_loss_weight: 1.0
49
+ rkd_loss_weight: 1.0
50
+ rkd_distance_weight: 1.0
51
+ rkd_angle_weight: 2.0
52
+
53
+ seed: 42
54
+ # Single GPU/process, so this is also the global batch size.
55
+ batch_size: 64
56
+ steps: 60000
57
+ log_freq: 200
58
+ output_dir: ${HOME}/groot_insight/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
59
+ job_name: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
60
+
61
+ wandb:
62
+ enable: true
63
+ disable_artifact: true
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,361 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 0.0,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 1e-05,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "initial_lr": 0.0001,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68,
88
+ 69,
89
+ 70,
90
+ 71,
91
+ 72,
92
+ 73,
93
+ 74,
94
+ 75,
95
+ 76,
96
+ 77,
97
+ 78,
98
+ 79,
99
+ 80,
100
+ 81,
101
+ 82,
102
+ 83,
103
+ 84,
104
+ 85,
105
+ 86,
106
+ 87,
107
+ 88,
108
+ 89,
109
+ 90,
110
+ 91,
111
+ 92,
112
+ 93,
113
+ 94,
114
+ 95,
115
+ 96,
116
+ 97,
117
+ 98,
118
+ 99,
119
+ 100,
120
+ 101,
121
+ 102,
122
+ 103,
123
+ 104,
124
+ 105,
125
+ 106,
126
+ 107,
127
+ 108,
128
+ 109,
129
+ 110,
130
+ 111,
131
+ 112,
132
+ 113,
133
+ 114,
134
+ 115,
135
+ 116,
136
+ 117,
137
+ 118,
138
+ 119,
139
+ 120,
140
+ 121,
141
+ 122,
142
+ 123,
143
+ 124,
144
+ 125,
145
+ 126,
146
+ 127,
147
+ 128,
148
+ 129,
149
+ 130,
150
+ 131,
151
+ 132,
152
+ 133,
153
+ 134,
154
+ 135,
155
+ 136,
156
+ 137,
157
+ 138,
158
+ 139,
159
+ 140,
160
+ 141,
161
+ 142,
162
+ 143,
163
+ 144,
164
+ 145,
165
+ 146,
166
+ 147,
167
+ 148,
168
+ 149,
169
+ 150,
170
+ 151,
171
+ 152,
172
+ 153,
173
+ 154,
174
+ 155,
175
+ 156,
176
+ 157,
177
+ 158
178
+ ]
179
+ },
180
+ {
181
+ "weight_decay": 0.0,
182
+ "lr": 0.0,
183
+ "betas": [
184
+ 0.9,
185
+ 0.999
186
+ ],
187
+ "eps": 1e-08,
188
+ "amsgrad": false,
189
+ "maximize": false,
190
+ "foreach": null,
191
+ "capturable": false,
192
+ "differentiable": false,
193
+ "fused": null,
194
+ "decoupled_weight_decay": true,
195
+ "initial_lr": 0.0001,
196
+ "params": [
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267,
306
+ 268,
307
+ 269,
308
+ 270,
309
+ 271,
310
+ 272,
311
+ 273,
312
+ 274,
313
+ 275,
314
+ 276,
315
+ 277,
316
+ 278,
317
+ 279,
318
+ 280,
319
+ 281,
320
+ 282,
321
+ 283,
322
+ 284,
323
+ 285,
324
+ 286,
325
+ 287,
326
+ 288,
327
+ 289,
328
+ 290,
329
+ 291,
330
+ 292,
331
+ 293,
332
+ 294,
333
+ 295,
334
+ 296,
335
+ 297,
336
+ 298,
337
+ 299,
338
+ 300,
339
+ 301,
340
+ 302,
341
+ 303,
342
+ 304,
343
+ 305,
344
+ 306,
345
+ 307,
346
+ 308,
347
+ 309,
348
+ 310,
349
+ 311,
350
+ 312,
351
+ 313,
352
+ 314,
353
+ 315,
354
+ 316,
355
+ 317,
356
+ 318,
357
+ 319,
358
+ 320
359
+ ]
360
+ }
361
+ ]
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/rscl_gpu1_rolling5k_20260723_022922.log ADDED
@@ -0,0 +1,198 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ INFO 2026-07-23 02:29:37 ot_train.py:281 {'batch_size': 64,
2
+ 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000',
3
+ 'cudnn_deterministic': False,
4
+ 'dataset': {'depth_output_unit': 'mm',
5
+ 'episodes': None,
6
+ 'eval_split': 0.0,
7
+ 'image_transforms': {'enable': False,
8
+ 'max_num_transforms': 3,
9
+ 'random_order': False,
10
+ 'tfs': {'affine': {'kwargs': {'degrees': [-5.0,
11
+ 5.0],
12
+ 'translate': [0.05,
13
+ 0.05]},
14
+ 'type': 'RandomAffine',
15
+ 'weight': 1.0},
16
+ 'brightness': {'kwargs': {'brightness': [0.8,
17
+ 1.2]},
18
+ 'type': 'ColorJitter',
19
+ 'weight': 1.0},
20
+ 'contrast': {'kwargs': {'contrast': [0.8,
21
+ 1.2]},
22
+ 'type': 'ColorJitter',
23
+ 'weight': 1.0},
24
+ 'hue': {'kwargs': {'hue': [-0.05,
25
+ 0.05]},
26
+ 'type': 'ColorJitter',
27
+ 'weight': 1.0},
28
+ 'saturation': {'kwargs': {'saturation': [0.5,
29
+ 1.5]},
30
+ 'type': 'ColorJitter',
31
+ 'weight': 1.0},
32
+ 'sharpness': {'kwargs': {'sharpness': [0.5,
33
+ 1.5]},
34
+ 'type': 'SharpnessJitter',
35
+ 'weight': 1.0}}},
36
+ 'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2',
37
+ 'return_uint8': False,
38
+ 'revision': None,
39
+ 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2',
40
+ 'streaming': False,
41
+ 'use_imagenet_stats': True,
42
+ 'video_backend': 'torchcodec'},
43
+ 'env': None,
44
+ 'env_eval_freq': 20000,
45
+ 'eval': {'batch_size': 50,
46
+ 'n_episodes': 50,
47
+ 'recording': False,
48
+ 'recording_private': False,
49
+ 'recording_repo_id': None,
50
+ 'use_async_envs': True},
51
+ 'eval_steps': 0,
52
+ 'job': {'detach': False,
53
+ 'image': 'huggingface/lerobot-gpu:latest',
54
+ 'tags': [],
55
+ 'target': None,
56
+ 'timeout': '2d'},
57
+ 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60',
58
+ 'log_freq': 200,
59
+ 'max_eval_samples': 0,
60
+ 'num_workers': 4,
61
+ 'optimizer': {'betas': [0.9, 0.999],
62
+ 'eps': 1e-08,
63
+ 'grad_clip_norm': 1.0,
64
+ 'lr': 0.0001,
65
+ 'type': 'adamw',
66
+ 'weight_decay': 1e-05},
67
+ 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K',
68
+ 'peft': None,
69
+ 'persistent_workers': True,
70
+ 'policy': {'action_decode_transform': None,
71
+ 'action_expert_num_layers': 16,
72
+ 'balance_dataset_weights': True,
73
+ 'balance_trajectory_weights': True,
74
+ 'base_model_path': 'nvidia/GR00T-N1.7-3B',
75
+ 'batch_size': 32,
76
+ 'chunk_size': 16,
77
+ 'dataloader_num_workers': 8,
78
+ 'dataset_paths': None,
79
+ 'device': 'cuda',
80
+ 'embodiment_tag': 'new_embodiment',
81
+ 'image_size': [256, 256],
82
+ 'input_features': {'observation.images.guide': {'shape': [3,
83
+ 224,
84
+ 224],
85
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
86
+ 'observation.images.right_shoulder': {'shape': [3,
87
+ 224,
88
+ 224],
89
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
90
+ 'observation.images.wrist': {'shape': [3,
91
+ 224,
92
+ 224],
93
+ 'type': <FeatureType.VISUAL: 'VISUAL'>},
94
+ 'observation.state': {'shape': [16],
95
+ 'type': <FeatureType.STATE: 'STATE'>}},
96
+ 'license': None,
97
+ 'lora_alpha': 16,
98
+ 'lora_dropout': 0.1,
99
+ 'lora_full_model': False,
100
+ 'lora_rank': 0,
101
+ 'max_action_dim': 132,
102
+ 'max_state_dim': 132,
103
+ 'max_steps': 10000,
104
+ 'model_params_fp32': True,
105
+ 'n_action_steps': 16,
106
+ 'n_obs_steps': 1,
107
+ 'normalization_mapping': {'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>,
108
+ 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>,
109
+ 'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>},
110
+ 'num_inference_timesteps': None,
111
+ 'optimizer_betas': [0.9, 0.999],
112
+ 'optimizer_eps': 1e-08,
113
+ 'optimizer_lr': 0.0001,
114
+ 'optimizer_weight_decay': 1e-05,
115
+ 'output_dir': './tmp/gr00t',
116
+ 'output_features': {'action': {'shape': [10],
117
+ 'type': <FeatureType.ACTION: 'ACTION'>}},
118
+ 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model',
119
+ 'pretrained_revision': None,
120
+ 'private': None,
121
+ 'push_to_hub': False,
122
+ 'relative_exclude_joints': [],
123
+ 'repo_id': None,
124
+ 'report_to': 'wandb',
125
+ 'resume': False,
126
+ 'rkd_angle_weight': 2.0,
127
+ 'rkd_distance_weight': 1.0,
128
+ 'rkd_enabled': True,
129
+ 'rkd_fm_loss_weight': 1.0,
130
+ 'rkd_loss_weight': 1.0,
131
+ 'rkd_student_attention_hidden_dim': 512,
132
+ 'rkd_student_attention_output_dim': 512,
133
+ 'rkd_student_pooling': 'attention',
134
+ 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11',
135
+ 'rtc_ramp_rate': None,
136
+ 'save_steps': 1000,
137
+ 'tags': None,
138
+ 'tokenizer_assets_repo': None,
139
+ 'tune_diffusion_model': True,
140
+ 'tune_llm': False,
141
+ 'tune_projector': True,
142
+ 'tune_top_llm_layers': 0,
143
+ 'tune_visual': False,
144
+ 'tune_vlln': True,
145
+ 'type': 'groot_rkd',
146
+ 'use_amp': False,
147
+ 'use_bf16': True,
148
+ 'use_flash_attention': False,
149
+ 'use_peft': False,
150
+ 'use_relative_actions': False,
151
+ 'video_backend': 'decord',
152
+ 'warmup_ratio': 0.05,
153
+ 'warmup_steps': 3000},
154
+ 'prefetch_factor': 4,
155
+ 'rename_map': {},
156
+ 'resume': True,
157
+ 'reward_model': None,
158
+ 'sample_weighting': None,
159
+ 'save_checkpoint': True,
160
+ 'save_checkpoint_to_hub': False,
161
+ 'save_freq': 20000,
162
+ 'scheduler': {'name': 'cosine', 'num_warmup_steps': 3000, 'type': 'diffuser'},
163
+ 'seed': 42,
164
+ 'steps': 60000,
165
+ 'tolerance_s': 0.0001,
166
+ 'use_policy_training_preset': True,
167
+ 'wandb': {'add_tags': True,
168
+ 'disable_artifact': True,
169
+ 'enable': True,
170
+ 'entity': None,
171
+ 'mode': None,
172
+ 'notes': None,
173
+ 'project': 'lerobot',
174
+ 'run_id': 'cemf0rge'}}
175
+ Traceback (most recent call last):
176
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1613, in <module>
177
+ main()
178
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1609, in main
179
+ train_module.main()
180
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 825, in main
181
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/configs/parser.py", line 320, in wrapper_inner
182
+ response = fn(cfg, *args, **kwargs)
183
+ ^^^^^^^^^^^^^^^^^^^^^^^^
184
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 285, in train
185
+ if not policy.is_trainable:
186
+ ^^^^^^^^^^^^^
187
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/common/wandb_utils.py", line 99, in __init__
188
+ wandb.init(
189
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 1528, in init
190
+ get_sentry().reraise(e)
191
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/analytics/sentry.py", line 191, in reraise
192
+ raise exc.with_traceback(tb)
193
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 1511, in init
194
+ run = wi.init(run_settings, run_config, run_printer)
195
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
196
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/wandb_init.py", line 961, in init
197
+ raise error
198
+ wandb.errors.errors.UsageError: You provided an invalid value for the `resume` argument. The value 'must' is not a valid option for resuming the run (cemf0rge) that has not been initialized. Please check your inputs and try again with a valid run ID. If you are trying to start a new run, please omit the `resume` argument or use `resume='allow'`.
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/rscl_gpu1_rolling5k_20260723_023110_newwandb.log ADDED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/source_config.yaml CHANGED
@@ -1,6 +1,8 @@
1
  # INSIGHT GR00T N1.7 RKD training with RS-CL-style full-training regularization.
2
  # Common RKD values live in the top-level policy; the phase contains schedule overrides only.
3
  insight:
 
 
4
  # Options: joint_abs | ee_abs_quat | ee_abs_rot6d | ee_delta
5
  action_space: ee_abs_rot6d
6
  # Options: coarse | detailed
 
1
  # INSIGHT GR00T N1.7 RKD training with RS-CL-style full-training regularization.
2
  # Common RKD values live in the top-level policy; the phase contains schedule overrides only.
3
  insight:
4
+ checkpoint_interval: 5K
5
+ checkpoint_slots: 2
6
  # Options: joint_abs | ee_abs_quat | ee_abs_rot6d | ee_delta
7
  action_space: ee_abs_rot6d
8
  # Options: coarse | detailed
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/debug-internal.log CHANGED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/debug.log CHANGED
@@ -1,20 +1,65 @@
1
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_setup.py:_flush():81] Current SDK version is 0.27.2
2
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_setup.py:_flush():81] Configure stats pid to 814719
3
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug.log
5
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug-internal.log
6
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_init.py:init():766] calling init triggers
7
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_init.py:init():771] wandb.init called with sweep_config: {}
8
- config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2', 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2', 'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 1.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 1.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'torchcodec', 'return_uint8': False, 'depth_output_unit': 'mm', 'streaming': False, 'eval_split': 0.0}, 'env': None, 'policy': {'type': 'groot_rkd', 'n_obs_steps': 1, 'input_features': {'observation.state': {'type': <FeatureType.STATE: 'STATE'>, 'shape': [16]}, 'observation.images.wrist': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.right_shoulder': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.guide': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}}, 'output_features': {'action': {'type': <FeatureType.ACTION: 'ACTION'>, 'shape': [10]}}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': None, 'private': None, 'tags': None, 'license': None, 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model', 'pretrained_revision': None, 'chunk_size': 16, 'n_action_steps': 16, 'max_state_dim': 132, 'max_action_dim': 132, 'normalization_mapping': {'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>}, 'base_model_path': 'nvidia/GR00T-N1.7-3B', 'action_decode_transform': None, 'embodiment_tag': 'new_embodiment', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'tune_vlln': True, 'tune_top_llm_layers': 0, 'num_inference_timesteps': None, 'rtc_ramp_rate': None, 'use_flash_attention': False, 'use_relative_actions': False, 'relative_exclude_joints': [], 'optimizer_lr': 0.0001, 'optimizer_betas': [0.9, 0.999], 'optimizer_eps': 1e-08, 'optimizer_weight_decay': 1e-05, 'warmup_steps': 3000, 'warmup_ratio': 0.05, 'use_bf16': True, 'model_params_fp32': True, 'image_size': [256, 256], 'tokenizer_assets_repo': None, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_full_model': False, 'video_backend': 'decord', 'balance_dataset_weights': True, 'balance_trajectory_weights': True, 'dataset_paths': None, 'output_dir': './tmp/gr00t', 'save_steps': 1000, 'max_steps': 10000, 'batch_size': 32, 'dataloader_num_workers': 8, 'report_to': 'wandb', 'resume': False, 'action_expert_num_layers': 16, 'rkd_enabled': True, 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11', 'rkd_student_pooling': 'attention', 'rkd_student_attention_hidden_dim': 512, 'rkd_student_attention_output_dim': 512, 'rkd_fm_loss_weight': 1.0, 'rkd_loss_weight': 1.0, 'rkd_distance_weight': 1.0, 'rkd_angle_weight': 2.0}, 'reward_model': None, 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K', 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60', 'resume': True, 'seed': 42, 'cudnn_deterministic': False, 'num_workers': 4, 'batch_size': 64, 'prefetch_factor': 4, 'persistent_workers': True, 'steps': 60000, 'env_eval_freq': 20000, 'log_freq': 200, 'eval_steps': 0, 'max_eval_samples': 0, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 20000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 0.0001, 'weight_decay': 1e-05, 'grad_clip_norm': 1.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': {'type': 'diffuser', 'num_warmup_steps': 3000, 'name': 'cosine'}, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': True, 'recording': False, 'recording_repo_id': None, 'recording_private': False}, 'wandb': {'enable': True, 'disable_artifact': True, 'project': 'lerobot', 'entity': None, 'notes': None, 'run_id': 'pgr3eari', 'mode': None, 'add_tags': True}, 'peft': None, 'job': {'target': None, 'image': 'huggingface/lerobot-gpu:latest', 'timeout': '2d', 'detach': False, 'tags': []}, 'save_checkpoint_to_hub': False, 'sample_weighting': None, 'rename_map': {}, 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000', '_wandb': {}}
9
- 2026-07-23 01:33:42,299 INFO MainThread:814719 [wandb_init.py:init():814] starting backend
10
- 2026-07-23 01:33:42,517 INFO MainThread:814719 [wandb_init.py:init():829] sending inform_init request
11
- 2026-07-23 01:33:42,889 INFO MainThread:814719 [wandb_init.py:init():834] backend started and connected
12
- 2026-07-23 01:33:42,891 INFO MainThread:814719 [wandb_init.py:init():904] updated telemetry
13
- 2026-07-23 01:33:42,896 INFO MainThread:814719 [wandb_init.py:init():927] communicating run to backend with 90.0 second timeout
14
- 2026-07-23 01:33:43,638 INFO MainThread:814719 [wandb_init.py:init():967] run resumed
15
- 2026-07-23 01:33:43,640 INFO MainThread:814719 [wandb_init.py:init():972] starting run threads in backend
16
- 2026-07-23 01:33:43,704 INFO MainThread:814719 [wandb_run.py:_console_start():2523] atexit reg
17
- 2026-07-23 01:33:43,704 INFO MainThread:814719 [wandb_run.py:_redirect():2373] redirect: wrap_raw
18
- 2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2442] Wrapping output streams.
19
- 2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2465] Redirects installed.
20
- 2026-07-23 01:33:43,706 INFO MainThread:814719 [wandb_init.py:init():1010] run started, returning control to user process
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Current SDK version is 0.27.2
2
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Configure stats pid to 1459
3
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug.log
5
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-internal.log
6
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():766] calling init triggers
7
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():771] wandb.init called with sweep_config: {}
8
+ config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2', 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2', 'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 1.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 1.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'torchcodec', 'return_uint8': False, 'depth_output_unit': 'mm', 'streaming': False, 'eval_split': 0.0}, 'env': None, 'policy': {'type': 'groot_rkd', 'n_obs_steps': 1, 'input_features': {'observation.state': {'type': <FeatureType.STATE: 'STATE'>, 'shape': [16]}, 'observation.images.wrist': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.right_shoulder': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.guide': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}}, 'output_features': {'action': {'type': <FeatureType.ACTION: 'ACTION'>, 'shape': [10]}}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': None, 'private': None, 'tags': None, 'license': None, 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model', 'pretrained_revision': None, 'chunk_size': 16, 'n_action_steps': 16, 'max_state_dim': 132, 'max_action_dim': 132, 'normalization_mapping': {'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>}, 'base_model_path': 'nvidia/GR00T-N1.7-3B', 'action_decode_transform': None, 'embodiment_tag': 'new_embodiment', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'tune_vlln': True, 'tune_top_llm_layers': 0, 'num_inference_timesteps': None, 'rtc_ramp_rate': None, 'use_flash_attention': False, 'use_relative_actions': False, 'relative_exclude_joints': [], 'optimizer_lr': 0.0001, 'optimizer_betas': [0.9, 0.999], 'optimizer_eps': 1e-08, 'optimizer_weight_decay': 1e-05, 'warmup_steps': 3000, 'warmup_ratio': 0.05, 'use_bf16': True, 'model_params_fp32': True, 'image_size': [256, 256], 'tokenizer_assets_repo': None, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_full_model': False, 'video_backend': 'decord', 'balance_dataset_weights': True, 'balance_trajectory_weights': True, 'dataset_paths': None, 'output_dir': './tmp/gr00t', 'save_steps': 1000, 'max_steps': 10000, 'batch_size': 32, 'dataloader_num_workers': 8, 'report_to': 'wandb', 'resume': False, 'action_expert_num_layers': 16, 'rkd_enabled': True, 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11', 'rkd_student_pooling': 'attention', 'rkd_student_attention_hidden_dim': 512, 'rkd_student_attention_output_dim': 512, 'rkd_fm_loss_weight': 1.0, 'rkd_loss_weight': 1.0, 'rkd_distance_weight': 1.0, 'rkd_angle_weight': 2.0}, 'reward_model': None, 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K', 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60', 'resume': True, 'seed': 42, 'cudnn_deterministic': False, 'num_workers': 4, 'batch_size': 64, 'prefetch_factor': 4, 'persistent_workers': True, 'steps': 60000, 'env_eval_freq': 20000, 'log_freq': 200, 'eval_steps': 0, 'max_eval_samples': 0, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 20000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 0.0001, 'weight_decay': 1e-05, 'grad_clip_norm': 1.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': {'type': 'diffuser', 'num_warmup_steps': 3000, 'name': 'cosine'}, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': True, 'recording': False, 'recording_repo_id': None, 'recording_private': False}, 'wandb': {'enable': True, 'disable_artifact': True, 'project': 'lerobot', 'entity': None, 'notes': None, 'run_id': 'rs1egwv8', 'mode': None, 'add_tags': True}, 'peft': None, 'job': {'target': None, 'image': 'huggingface/lerobot-gpu:latest', 'timeout': '2d', 'detach': False, 'tags': []}, 'save_checkpoint_to_hub': False, 'sample_weighting': None, 'rename_map': {}, 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01', '_wandb': {}}
9
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():814] starting backend
10
+ 2026-07-23 16:35:01,833 INFO MainThread:1459 [wandb_init.py:init():829] sending inform_init request
11
+ 2026-07-23 16:35:02,323 INFO MainThread:1459 [wandb_init.py:init():834] backend started and connected
12
+ 2026-07-23 16:35:02,325 INFO MainThread:1459 [wandb_init.py:init():904] updated telemetry
13
+ 2026-07-23 16:35:02,326 INFO MainThread:1459 [wandb_init.py:init():927] communicating run to backend with 90.0 second timeout
14
+ 2026-07-23 16:35:03,342 INFO MainThread:1459 [wandb_init.py:init():967] run resumed
15
+ 2026-07-23 16:35:03,344 INFO MainThread:1459 [wandb_init.py:init():972] starting run threads in backend
16
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_console_start():2523] atexit reg
17
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2373] redirect: wrap_raw
18
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2442] Wrapping output streams.
19
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2465] Redirects installed.
20
+ 2026-07-23 16:35:03,398 INFO MainThread:1459 [wandb_init.py:init():1010] run started, returning control to user process
21
+ 2026-07-24 04:10:06,982 INFO wandb-AsyncioManager-main:1459 [service_client.py:_forward_responses():122] Reached EOF.
22
+ 2026-07-24 04:10:06,983 INFO wandb-AsyncioManager-main:1459 [mailbox.py:close():154] Closing mailbox, abandoning 2 handles.
23
+ 2026-07-24 04:10:06,983 ERROR wandb-AsyncioManager-main:1459 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
24
+ Traceback (most recent call last):
25
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
26
+ await fn()
27
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 90, in loop
28
+ await asyncio_compat.race(
29
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 278, in race
30
+ async with open_task_group(race=True) as tg:
31
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^
32
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/contextlib.py", line 217, in __aexit__
33
+ await anext(self.gen)
34
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 240, in open_task_group
35
+ await task_group._wait_all(race=race, timeout=exit_timeout)
36
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 180, in _wait_all
37
+ raise exc
38
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 113, in _print_all
39
+ result = await handle.wait_async(timeout=None)
40
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
41
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox_handle.py", line 127, in wait_async
42
+ response = await self._handle.wait_async(timeout=timeout)
43
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
44
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/response_handle.py", line 123, in wait_async
45
+ raise HandleAbandonedError()
46
+ wandb.sdk.mailbox.mailbox_handle.HandleAbandonedError
47
+ 2026-07-24 04:10:14,759 ERROR wandb-AsyncioManager-main:1459 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
48
+ Traceback (most recent call last):
49
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
50
+ await fn()
51
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 87, in loop
52
+ should_stop = await self._check_should_stop()
53
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
54
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 115, in _check_should_stop
55
+ handle = await self._interface.deliver_async(request)
56
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
57
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/interface/interface_sock.py", line 58, in deliver_async
58
+ handle = await self._client.deliver(request)
59
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
60
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 57, in deliver
61
+ handle = self._mailbox.require_response(request)
62
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
63
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox.py", line 89, in require_response
64
+ raise MailboxClosedError()
65
+ wandb.sdk.mailbox.mailbox.MailboxClosedError
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/config.yaml ADDED
@@ -0,0 +1,334 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ _wandb:
2
+ value:
3
+ cli_version: 0.27.2
4
+ e:
5
+ vmsr43xlhcpoc4wtuae54k6a709gs2et:
6
+ args:
7
+ - --config_path=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model/train_config.json
8
+ - --resume=true
9
+ - --output_dir=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
10
+ - --dataset.root=/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2
11
+ codePath: benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py
12
+ codePathLocal: my_scripts/lerobot_train_yaml.py
13
+ cpu_count: 112
14
+ cpu_count_logical: 224
15
+ cudaVersion: "12.8"
16
+ disk:
17
+ /:
18
+ total: "1888425144320"
19
+ used: "1137071321088"
20
+ executable: /home/ext_minje/miniconda3/envs/lerobot060_groot/bin/python
21
+ git:
22
+ commit: a329728c2c9524d79b966c4a7a82ee56a4c2abf7
23
+ remote: git@github.com:paragon7060/clvla.git
24
+ gpu: NVIDIA H200
25
+ gpu_count: 2
26
+ gpu_nvidia:
27
+ - architecture: Hopper
28
+ cudaCores: 16896
29
+ memoryTotal: "150754820096"
30
+ name: NVIDIA H200
31
+ uuid: GPU-86be6e5b-fb48-fa1c-9bd7-83d336cb2991
32
+ - architecture: Hopper
33
+ cudaCores: 16896
34
+ memoryTotal: "150754820096"
35
+ name: NVIDIA H200
36
+ uuid: GPU-24eb1ba2-7313-ff93-1435-748768ebdbdc
37
+ host: ext_minje-box
38
+ memory:
39
+ total: "2164176568320"
40
+ os: Linux-6.8.0-55-generic-x86_64-with-glibc2.39
41
+ program: /home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py
42
+ python: CPython 3.12.13
43
+ root: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
44
+ startedAt: "2026-07-23T01:33:42.298193Z"
45
+ writerId: vmsr43xlhcpoc4wtuae54k6a709gs2et
46
+ m: []
47
+ python_version: 3.12.13
48
+ t:
49
+ "1":
50
+ - 1
51
+ - 11
52
+ - 41
53
+ - 49
54
+ - 51
55
+ - 71
56
+ - 83
57
+ - 98
58
+ "2":
59
+ - 1
60
+ - 11
61
+ - 41
62
+ - 49
63
+ - 51
64
+ - 71
65
+ - 83
66
+ - 98
67
+ "3":
68
+ - 5
69
+ - 13
70
+ - 14
71
+ - 15
72
+ - 16
73
+ - 61
74
+ - 62
75
+ "4": 3.12.13
76
+ "5": 0.27.2
77
+ "6": 5.5.4
78
+ "10":
79
+ - 21
80
+ "12": 0.27.2
81
+ "13": linux-x86_64
82
+ batch_size:
83
+ value: 64
84
+ checkpoint_path:
85
+ value: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000
86
+ cudnn_deterministic:
87
+ value: false
88
+ dataset:
89
+ value:
90
+ depth_output_unit: mm
91
+ episodes: null
92
+ eval_split: 0
93
+ image_transforms:
94
+ enable: false
95
+ max_num_transforms: 3
96
+ random_order: false
97
+ tfs:
98
+ affine:
99
+ kwargs:
100
+ degrees:
101
+ - -5
102
+ - 5
103
+ translate:
104
+ - 0.05
105
+ - 0.05
106
+ type: RandomAffine
107
+ weight: 1
108
+ brightness:
109
+ kwargs:
110
+ brightness:
111
+ - 0.8
112
+ - 1.2
113
+ type: ColorJitter
114
+ weight: 1
115
+ contrast:
116
+ kwargs:
117
+ contrast:
118
+ - 0.8
119
+ - 1.2
120
+ type: ColorJitter
121
+ weight: 1
122
+ hue:
123
+ kwargs:
124
+ hue:
125
+ - -0.05
126
+ - 0.05
127
+ type: ColorJitter
128
+ weight: 1
129
+ saturation:
130
+ kwargs:
131
+ saturation:
132
+ - 0.5
133
+ - 1.5
134
+ type: ColorJitter
135
+ weight: 1
136
+ sharpness:
137
+ kwargs:
138
+ sharpness:
139
+ - 0.5
140
+ - 1.5
141
+ type: SharpnessJitter
142
+ weight: 1
143
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
144
+ return_uint8: false
145
+ revision: null
146
+ root: /home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2
147
+ streaming: false
148
+ use_imagenet_stats: true
149
+ video_backend: torchcodec
150
+ env:
151
+ value: null
152
+ env_eval_freq:
153
+ value: 20000
154
+ eval:
155
+ value:
156
+ batch_size: 50
157
+ n_episodes: 50
158
+ recording: false
159
+ recording_private: false
160
+ recording_repo_id: null
161
+ use_async_envs: true
162
+ eval_steps:
163
+ value: 0
164
+ job:
165
+ value:
166
+ detach: false
167
+ image: huggingface/lerobot-gpu:latest
168
+ tags: []
169
+ target: null
170
+ timeout: 2d
171
+ job_name:
172
+ value: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
173
+ log_freq:
174
+ value: 200
175
+ max_eval_samples:
176
+ value: 0
177
+ num_workers:
178
+ value: 4
179
+ optimizer:
180
+ value:
181
+ betas:
182
+ - 0.9
183
+ - 0.999
184
+ eps: 1e-08
185
+ grad_clip_norm: 1
186
+ lr: 0.0001
187
+ type: adamw
188
+ weight_decay: 1e-05
189
+ output_dir:
190
+ value: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
191
+ peft:
192
+ value: null
193
+ persistent_workers:
194
+ value: true
195
+ policy:
196
+ value:
197
+ action_decode_transform: null
198
+ action_expert_num_layers: 16
199
+ balance_dataset_weights: true
200
+ balance_trajectory_weights: true
201
+ base_model_path: nvidia/GR00T-N1.7-3B
202
+ batch_size: 32
203
+ chunk_size: 16
204
+ dataloader_num_workers: 8
205
+ dataset_paths: null
206
+ device: cuda
207
+ embodiment_tag: new_embodiment
208
+ image_size:
209
+ - 256
210
+ - 256
211
+ input_features:
212
+ observation.images.guide:
213
+ shape:
214
+ - 3
215
+ - 224
216
+ - 224
217
+ type: VISUAL
218
+ observation.images.right_shoulder:
219
+ shape:
220
+ - 3
221
+ - 224
222
+ - 224
223
+ type: VISUAL
224
+ observation.images.wrist:
225
+ shape:
226
+ - 3
227
+ - 224
228
+ - 224
229
+ type: VISUAL
230
+ observation.state:
231
+ shape:
232
+ - 16
233
+ type: STATE
234
+ license: null
235
+ lora_alpha: 16
236
+ lora_dropout: 0.1
237
+ lora_full_model: false
238
+ lora_rank: 0
239
+ max_action_dim: 132
240
+ max_state_dim: 132
241
+ max_steps: 10000
242
+ model_params_fp32: true
243
+ n_action_steps: 16
244
+ n_obs_steps: 1
245
+ normalization_mapping:
246
+ ACTION: IDENTITY
247
+ STATE: IDENTITY
248
+ VISUAL: IDENTITY
249
+ num_inference_timesteps: null
250
+ optimizer_betas:
251
+ - 0.9
252
+ - 0.999
253
+ optimizer_eps: 1e-08
254
+ optimizer_lr: 0.0001
255
+ optimizer_weight_decay: 1e-05
256
+ output_dir: ./tmp/gr00t
257
+ output_features:
258
+ action:
259
+ shape:
260
+ - 10
261
+ type: ACTION
262
+ pretrained_path: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model
263
+ pretrained_revision: null
264
+ private: null
265
+ push_to_hub: false
266
+ relative_exclude_joints: []
267
+ repo_id: null
268
+ report_to: wandb
269
+ resume: false
270
+ rkd_angle_weight: 2
271
+ rkd_distance_weight: 1
272
+ rkd_enabled: true
273
+ rkd_fm_loss_weight: 1
274
+ rkd_loss_weight: 1
275
+ rkd_student_attention_hidden_dim: 512
276
+ rkd_student_attention_output_dim: 512
277
+ rkd_student_pooling: attention
278
+ rkd_vae_checkpoint: /home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
279
+ rtc_ramp_rate: null
280
+ save_steps: 1000
281
+ tags: null
282
+ tokenizer_assets_repo: null
283
+ tune_diffusion_model: true
284
+ tune_llm: false
285
+ tune_projector: true
286
+ tune_top_llm_layers: 0
287
+ tune_visual: false
288
+ tune_vlln: true
289
+ type: groot_rkd
290
+ use_amp: false
291
+ use_bf16: true
292
+ use_flash_attention: false
293
+ use_peft: false
294
+ use_relative_actions: false
295
+ video_backend: decord
296
+ warmup_ratio: 0.05
297
+ warmup_steps: 3000
298
+ prefetch_factor:
299
+ value: 4
300
+ resume:
301
+ value: true
302
+ reward_model:
303
+ value: null
304
+ sample_weighting:
305
+ value: null
306
+ save_checkpoint:
307
+ value: true
308
+ save_checkpoint_to_hub:
309
+ value: false
310
+ save_freq:
311
+ value: 20000
312
+ scheduler:
313
+ value:
314
+ name: cosine
315
+ num_warmup_steps: 3000
316
+ type: diffuser
317
+ seed:
318
+ value: 42
319
+ steps:
320
+ value: 60000
321
+ tolerance_s:
322
+ value: 0.0001
323
+ use_policy_training_preset:
324
+ value: true
325
+ wandb:
326
+ value:
327
+ add_tags: true
328
+ disable_artifact: true
329
+ enable: true
330
+ entity: null
331
+ mode: null
332
+ notes: null
333
+ project: lerobot
334
+ run_id: pgr3eari
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/output.log CHANGED
@@ -45,4 +45,47 @@ Training: 3%|▎ | 800/30000 [20:31<12:21:37, 1.52s/step]INFO 2026-07
45
  Training: 3%|▎ | 1000/30000 [26:18<22:10:00, 2.75s/step]INFO 2026-07-23 02:00:30 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.09 loss:0.039 grdn:0.298 lr:5.2e-05 updt_s:0.843 data_s:0.886 smp/s:37 mem_gb:29.84
46
  Training: 4%|▍ | 1200/30000 [31:32<10:05:29, 1.26s/step]INFO 2026-07-23 02:05:45 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.14 loss:0.039 grdn:0.306 lr:5.1e-05 updt_s:0.805 data_s:0.765 smp/s:41 mem_gb:29.84
47
  Training: 5%|▍ | 1400/30000 [36:06<10:52:36, 1.37s/step]INFO 2026-07-23 02:10:19 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.20 loss:0.039 grdn:0.298 lr:5.1e-05 updt_s:0.744 data_s:0.622 smp/s:47 mem_gb:29.84
48
- Training: 5%| | 1482/30000 [38:04<10:54:48, 1.38s/step]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45
  Training: 3%|▎ | 1000/30000 [26:18<22:10:00, 2.75s/step]INFO 2026-07-23 02:00:30 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.09 loss:0.039 grdn:0.298 lr:5.2e-05 updt_s:0.843 data_s:0.886 smp/s:37 mem_gb:29.84
46
  Training: 4%|▍ | 1200/30000 [31:32<10:05:29, 1.26s/step]INFO 2026-07-23 02:05:45 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.14 loss:0.039 grdn:0.306 lr:5.1e-05 updt_s:0.805 data_s:0.765 smp/s:41 mem_gb:29.84
47
  Training: 5%|▍ | 1400/30000 [36:06<10:52:36, 1.37s/step]INFO 2026-07-23 02:10:19 ot_train.py:649 step:31K smpl:2M ep:45K epch:9.20 loss:0.039 grdn:0.298 lr:5.1e-05 updt_s:0.744 data_s:0.622 smp/s:47 mem_gb:29.84
48
+ Training: 5%| | 1600/30000 [41:01<19:58:58, 2.53s/step]INFO 2026-07-23 02:15:14 ot_train.py:649 step:32K smpl:2M ep:46K epch:9.26 loss:0.039 grdn:0.299 lr:5.0e-05 updt_s:0.806 data_s:0.666 smp/s:43 mem_gb:29.82
49
+ Training: 6%|▌ | 1800/30000 [45:51<10:43:33, 1.37s/step]INFO 2026-07-23 02:20:04 ot_train.py:649 step:32K smpl:2M ep:46K epch:9.32 loss:0.037 grdn:0.292 lr:4.9e-05 updt_s:0.787 data_s:0.659 smp/s:44 mem_gb:29.85
50
+ Training: 7%|▋ | 2000/30000 [50:39<10:41:47, 1.38s/step]INFO 2026-07-23 02:24:51 ot_train.py:649 step:32K smpl:2M ep:46K epch:9.38 loss:0.037 grdn:0.286 lr:4.9e-05 updt_s:0.776 data_s:0.660 smp/s:45 mem_gb:29.85
51
+ Training: 7%|▋ | 2167/30000 [54:37<10:34:57, 1.37s/step]Traceback (most recent call last):
52
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1453, in <module>
53
+ config_path = Path(config_value).expanduser()
54
+ ^^^^^^
55
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py", line 1449, in main
56
+ config_value = _find_cli_value(args, "--config_path")
57
+ ^^^^^^^^^^^^^^^^^^^
58
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 816, in main
59
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/configs/parser.py", line 320, in wrapper_inner
60
+ response = fn(cfg, *args, **kwargs)
61
+ ^^^^^^^^^^^^^^^^^^^^^^^^
62
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 618, in train
63
+ if is_eval_step:
64
+
65
+ File "/home/ext_minje/clvla/lerobot/src/lerobot/scripts/lerobot_train.py", line 179, in update_policy
66
+ """
67
+
68
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_policies/groot_RKD/modeling_groot_rkd.py", line 326, in forward
69
+ outputs = self._groot_model.forward(groot_inputs)
70
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
71
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_policies/groot_RKD/groot_n1_7.py", line 884, in forward
72
+ backbone_outputs = self.backbone(backbone_inputs)
73
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
74
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl
75
+ return self._call_impl(*args, **kwargs)
76
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
77
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl
78
+ return forward_call(*args, **kwargs)
79
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
80
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_policies/groot_RKD/groot_n1_7.py", line 450, in forward
81
+ self._ensure_legacy_qwen3_position_ids(model_input)
82
+ File "/home/ext_minje/clvla/benchmarks/INSIGHT/my_policies/groot_RKD/groot_n1_7.py", line 402, in _ensure_legacy_qwen3_position_ids
83
+ position_ids = compute_3d_position_ids(
84
+ ^^^^^^^^^^^^^^^^^^^^^^^^
85
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 1228, in compute_3d_position_ids
86
+ position_ids, rope_deltas = self.get_rope_index(
87
+ ^^^^^^^^^^^^^^^^^^^^
88
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 1106, in get_rope_index
89
+ current_pos += text_len
90
+ ^^^^^^^^^^^
91
+ KeyboardInterrupt
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/files/wandb-summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"train/rkd_teacher_latent_norm":1,"train/weighted_rkd_loss":0.001385301467962563,"_timestamp":1.7847403914008305e+09,"train/training_phase_start_step":0,"train/rkd_loss_weight":0.3208404665513721,"_wandb":{"runtime":59006},"train/weighted_flow_matching_loss":0.022767670452594757,"train/episodes":53461.55783611993,"train/grad_norm":0.26664105713367464,"train/rkd_distance_loss":0.0010141097009181976,"train/training_phase_trainable_params":1.080755329e+09,"train/gpu_mem_gb":29.854091150760652,"train/flow_matching_loss":0.022767670452594757,"train/rkd_student_feature_norm":16.188758850097656,"train/rkd_runtime_enabled":1,"train/loss":0.024152971804142,"train/samples_per_s":49.78324781714454,"train/lr":3.5336255062138e-05,"train/steps":37000,"_step":37000,"train/rkd_angle_loss":0.0016518083866685629,"train/samples":2368000,"train/rkd_loss":0.004317726474255323,"train/training_phase_until_step":60000,"train/dataloading_s":0.5792085527597374,"_runtime":59006,"train/training_phase_index":0,"train/rkd_fm_loss_weight":1,"train/update_s":0.7063644623853179,"train/rkd_student_pooling_id":1,"train/epochs":10.844129378523313}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug-core.log CHANGED
@@ -26,3 +26,39 @@
26
  {"time":"2026-07-23T01:33:48.72926561Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"0uuooyt4of4j"}
27
  {"time":"2026-07-23T01:33:48.824026381Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"v9b9occvkv0z"}
28
  {"time":"2026-07-23T01:33:48.928044163Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"3mgxscqyzkyn"}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
  {"time":"2026-07-23T01:33:48.72926561Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"0uuooyt4of4j"}
27
  {"time":"2026-07-23T01:33:48.824026381Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"v9b9occvkv0z"}
28
  {"time":"2026-07-23T01:33:48.928044163Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"3mgxscqyzkyn"}
29
+ {"time":"2026-07-23T02:28:50.768484688Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
30
+ {"time":"2026-07-23T02:28:50.768650257Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
31
+ {"time":"2026-07-23T02:28:50.768660338Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
32
+ {"time":"2026-07-23T02:28:50.768702111Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
33
+ {"time":"2026-07-23T02:28:50.768725726Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
34
+ {"time":"2026-07-23T02:28:50.768737691Z","level":"INFO","msg":"server is shutting down"}
35
+ {"time":"2026-07-23T02:28:50.768728092Z","level":"INFO","msg":"server is shutting down"}
36
+ {"time":"2026-07-23T02:28:50.768751955Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
37
+ {"time":"2026-07-23T02:28:50.768745569Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
38
+ {"time":"2026-07-23T02:28:50.768805358Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
39
+ {"time":"2026-07-23T02:28:50.768815828Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-814707-815594-198914184/socket","Net":"unix"}}
40
+ {"time":"2026-07-23T02:28:50.76883043Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-814713-815525-1635480438/socket","Net":"unix"}}
41
+ {"time":"2026-07-23T02:28:50.770084447Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
42
+ {"time":"2026-07-23T02:28:50.770158191Z","level":"INFO","msg":"server is shutting down"}
43
+ {"time":"2026-07-23T02:28:50.770173329Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
44
+ {"time":"2026-07-23T02:28:50.770146502Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
45
+ {"time":"2026-07-23T02:28:50.770246375Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-814719-815551-3138903381/socket","Net":"unix"}}
46
+ {"time":"2026-07-23T02:28:50.770382899Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
47
+ {"time":"2026-07-23T02:28:50.773003008Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
48
+ {"time":"2026-07-23T02:28:50.773054953Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
49
+ {"time":"2026-07-23T02:28:50.773111864Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
50
+ {"time":"2026-07-23T02:28:50.773106808Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
51
+ {"time":"2026-07-23T02:28:50.773073579Z","level":"INFO","msg":"server is shutting down"}
52
+ {"time":"2026-07-23T02:28:50.773200905Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-814701-815561-915654594/socket","Net":"unix"}}
53
+ {"time":"2026-07-23T02:28:52.626908382Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
54
+ {"time":"2026-07-23T02:28:52.626929775Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
55
+ {"time":"2026-07-23T02:28:52.626940482Z","level":"INFO","msg":"server is closed"}
56
+ {"time":"2026-07-23T02:28:52.662753611Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
57
+ {"time":"2026-07-23T02:28:52.662779647Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
58
+ {"time":"2026-07-23T02:28:52.662788745Z","level":"INFO","msg":"server is closed"}
59
+ {"time":"2026-07-23T02:28:52.68387658Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
60
+ {"time":"2026-07-23T02:28:52.683943364Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
61
+ {"time":"2026-07-23T02:28:52.683953617Z","level":"INFO","msg":"server is closed"}
62
+ {"time":"2026-07-23T02:29:02.470789399Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
63
+ {"time":"2026-07-23T02:29:02.470826347Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
64
+ {"time":"2026-07-23T02:29:02.470833873Z","level":"INFO","msg":"server is closed"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug-internal.log CHANGED
@@ -245,3 +245,109 @@
245
  {"time":"2026-07-23T02:11:32.739406463Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":300,"events_lines":2,"console_offset":47,"console_lines":1}
246
  {"time":"2026-07-23T02:12:22.402029921Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
247
  {"time":"2026-07-23T02:12:22.4021656Z","level":"INFO","msg":"filestream: sending request","total_files":0}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
245
  {"time":"2026-07-23T02:11:32.739406463Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":300,"events_lines":2,"console_offset":47,"console_lines":1}
246
  {"time":"2026-07-23T02:12:22.402029921Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
247
  {"time":"2026-07-23T02:12:22.4021656Z","level":"INFO","msg":"filestream: sending request","total_files":0}
248
+ {"time":"2026-07-23T02:12:23.362164613Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
249
+ {"time":"2026-07-23T02:12:23.362395596Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":302,"events_lines":6,"console_offset":47,"console_lines":1}
250
+ {"time":"2026-07-23T02:12:24.144626493Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
251
+ {"time":"2026-07-23T02:12:24.144870858Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":47,"console_lines":1}
252
+ {"time":"2026-07-23T02:13:12.098678344Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
253
+ {"time":"2026-07-23T02:13:12.099003095Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":308,"events_lines":2,"console_offset":47,"console_lines":1}
254
+ {"time":"2026-07-23T02:13:12.458838875Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
255
+ {"time":"2026-07-23T02:13:12.459121578Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":310,"events_lines":4,"console_offset":47,"console_lines":1}
256
+ {"time":"2026-07-23T02:14:04.269679981Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
257
+ {"time":"2026-07-23T02:14:04.269842494Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":47,"console_lines":1}
258
+ {"time":"2026-07-23T02:14:04.959605431Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
259
+ {"time":"2026-07-23T02:14:04.959869932Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":314,"events_lines":8,"console_offset":47,"console_lines":1}
260
+ {"time":"2026-07-23T02:14:47.260819768Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
261
+ {"time":"2026-07-23T02:14:47.261088178Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":47,"console_lines":1}
262
+ {"time":"2026-07-23T02:14:47.816232955Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
263
+ {"time":"2026-07-23T02:14:47.816499019Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":322,"events_lines":6,"console_offset":47,"console_lines":1}
264
+ {"time":"2026-07-23T02:15:14.209850024Z","level":"WARN","msg":"handler: ignoring partial history record","step":31600,"current":37001}
265
+ {"time":"2026-07-23T02:15:23.521072135Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
266
+ {"time":"2026-07-23T02:15:23.521353534Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":47,"console_lines":1}
267
+ {"time":"2026-07-23T02:15:23.982612715Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
268
+ {"time":"2026-07-23T02:15:23.982826206Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":328,"events_lines":4,"console_offset":47,"console_lines":2}
269
+ {"time":"2026-07-23T02:15:24.425232875Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
270
+ {"time":"2026-07-23T02:15:24.425483013Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":48,"console_lines":1}
271
+ {"time":"2026-07-23T02:16:01.786406101Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
272
+ {"time":"2026-07-23T02:16:01.786603022Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":332,"events_lines":2,"console_offset":48,"console_lines":1}
273
+ {"time":"2026-07-23T02:16:25.713166877Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
274
+ {"time":"2026-07-23T02:16:25.713431331Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":334,"events_lines":4,"console_offset":48,"console_lines":1}
275
+ {"time":"2026-07-23T02:17:03.266808763Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
276
+ {"time":"2026-07-23T02:17:03.267087671Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":338,"events_lines":2,"console_offset":48,"console_lines":1}
277
+ {"time":"2026-07-23T02:17:50.530237524Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
278
+ {"time":"2026-07-23T02:17:50.530495713Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":340,"events_lines":6,"console_offset":48,"console_lines":1}
279
+ {"time":"2026-07-23T02:18:20.451787991Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
280
+ {"time":"2026-07-23T02:18:20.452026414Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":346,"events_lines":6,"console_offset":48,"console_lines":1}
281
+ {"time":"2026-07-23T02:19:01.788519665Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
282
+ {"time":"2026-07-23T02:19:01.788703845Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":352,"events_lines":4,"console_offset":48,"console_lines":1}
283
+ {"time":"2026-07-23T02:19:02.328045556Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
284
+ {"time":"2026-07-23T02:19:02.328268427Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":356,"events_lines":6,"console_offset":48,"console_lines":1}
285
+ {"time":"2026-07-23T02:19:02.759022028Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
286
+ {"time":"2026-07-23T02:19:02.759194577Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":48,"console_lines":1}
287
+ {"time":"2026-07-23T02:19:03.248435965Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
288
+ {"time":"2026-07-23T02:19:16.837240643Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":362,"events_lines":2,"console_offset":48,"console_lines":1}
289
+ {"time":"2026-07-23T02:19:17.303817335Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
290
+ {"time":"2026-07-23T02:19:31.836655244Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":364,"events_lines":2,"console_offset":48,"console_lines":1}
291
+ {"time":"2026-07-23T02:20:04.018094282Z","level":"WARN","msg":"handler: ignoring partial history record","step":31800,"current":37001}
292
+ {"time":"2026-07-23T02:20:13.742987022Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
293
+ {"time":"2026-07-23T02:20:13.743151354Z","level":"INFO","msg":"filestream: sending request","total_files":0}
294
+ {"time":"2026-07-23T02:20:14.186936411Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
295
+ {"time":"2026-07-23T02:20:14.187196098Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":366,"events_lines":5,"console_offset":48,"console_lines":2}
296
+ {"time":"2026-07-23T02:20:14.606163567Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
297
+ {"time":"2026-07-23T02:20:14.606358638Z","level":"INFO","msg":"filestream: sending request","total_files":1,"events_offset":371,"events_lines":1}
298
+ {"time":"2026-07-23T02:20:47.648537651Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
299
+ {"time":"2026-07-23T02:20:47.648719247Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":372,"events_lines":1,"console_offset":49,"console_lines":1}
300
+ {"time":"2026-07-23T02:20:48.12626836Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
301
+ {"time":"2026-07-23T02:20:48.126523688Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":373,"events_lines":3,"console_offset":49,"console_lines":1}
302
+ {"time":"2026-07-23T02:21:12.996289747Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
303
+ {"time":"2026-07-23T02:21:12.996538083Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":376,"events_lines":1,"console_offset":49,"console_lines":1}
304
+ {"time":"2026-07-23T02:21:40.076672259Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
305
+ {"time":"2026-07-23T02:21:40.076833315Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":377,"events_lines":2,"console_offset":49,"console_lines":1}
306
+ {"time":"2026-07-23T02:22:12.591294436Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
307
+ {"time":"2026-07-23T02:22:12.591534984Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":379,"events_lines":3,"console_offset":49,"console_lines":1}
308
+ {"time":"2026-07-23T02:22:13.008257666Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
309
+ {"time":"2026-07-23T02:22:13.008448544Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":382,"events_lines":4,"console_offset":49,"console_lines":1}
310
+ {"time":"2026-07-23T02:22:43.455358315Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
311
+ {"time":"2026-07-23T02:22:43.4555333Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":49,"console_lines":1}
312
+ {"time":"2026-07-23T02:23:15.177406156Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
313
+ {"time":"2026-07-23T02:23:15.177600701Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":386,"events_lines":4,"console_offset":49,"console_lines":1}
314
+ {"time":"2026-07-23T02:23:15.790816852Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
315
+ {"time":"2026-07-23T02:23:15.791043145Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":390,"events_lines":6,"console_offset":49,"console_lines":1}
316
+ {"time":"2026-07-23T02:23:16.511965589Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
317
+ {"time":"2026-07-23T02:23:16.531664799Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":49,"console_lines":1}
318
+ {"time":"2026-07-23T02:24:09.59999693Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
319
+ {"time":"2026-07-23T02:24:09.600190154Z","level":"INFO","msg":"filestream: sending request","total_files":0}
320
+ {"time":"2026-07-23T02:24:10.096383571Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
321
+ {"time":"2026-07-23T02:24:10.096565421Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":396,"events_lines":6,"console_offset":49,"console_lines":1}
322
+ {"time":"2026-07-23T02:24:10.56096879Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
323
+ {"time":"2026-07-23T02:24:10.87257939Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":49,"console_lines":1}
324
+ {"time":"2026-07-23T02:24:51.945277879Z","level":"WARN","msg":"handler: ignoring partial history record","step":32000,"current":37001}
325
+ {"time":"2026-07-23T02:24:53.714569624Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
326
+ {"time":"2026-07-23T02:24:53.714700687Z","level":"INFO","msg":"filestream: sending request","total_files":0}
327
+ {"time":"2026-07-23T02:24:54.093680458Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
328
+ {"time":"2026-07-23T02:24:54.093909985Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":402,"events_lines":6,"console_offset":49,"console_lines":2}
329
+ {"time":"2026-07-23T02:25:49.372929793Z","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/minje227_hyu-hanyang-university/lerobot/pgr3eari/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
330
+ {"time":"2026-07-23T02:26:24.414750595Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
331
+ {"time":"2026-07-23T02:26:24.414969006Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":50,"console_lines":1}
332
+ {"time":"2026-07-23T02:27:04.885943619Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
333
+ {"time":"2026-07-23T02:27:04.886281642Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":408,"events_lines":12,"console_offset":50,"console_lines":1}
334
+ {"time":"2026-07-23T02:27:05.35140455Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
335
+ {"time":"2026-07-23T02:27:05.351607052Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":420,"events_lines":6,"console_offset":50,"console_lines":1}
336
+ {"time":"2026-07-23T02:27:31.046145432Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
337
+ {"time":"2026-07-23T02:27:31.046284597Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":50,"console_lines":1}
338
+ {"time":"2026-07-23T02:27:58.5330972Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
339
+ {"time":"2026-07-23T02:27:58.533277753Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":426,"events_lines":4,"console_offset":50,"console_lines":1}
340
+ {"time":"2026-07-23T02:27:58.938590054Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
341
+ {"time":"2026-07-23T02:27:58.938897544Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":430,"events_lines":2,"console_offset":50,"console_lines":1}
342
+ {"time":"2026-07-23T02:28:32.127453304Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
343
+ {"time":"2026-07-23T02:28:32.127626571Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":432,"events_lines":2,"console_offset":50,"console_lines":1}
344
+ {"time":"2026-07-23T02:28:52.097937416Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
345
+ {"time":"2026-07-23T02:29:01.470938231Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
346
+ {"time":"2026-07-23T02:29:01.471166513Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":434,"events_lines":4,"console_offset":50,"console_lines":1}
347
+ {"time":"2026-07-23T02:29:01.947602559Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
348
+ {"time":"2026-07-23T02:29:01.947868498Z","level":"INFO","msg":"filestream: sending request","total_files":3,"events_offset":438,"events_lines":2,"console_offset":50,"console_lines":41,"uploaded_len":3,"complete":true,"exit_code":255}
349
+ {"time":"2026-07-23T02:29:02.470009043Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
350
+ {"time":"2026-07-23T02:29:02.470189702Z","level":"INFO","msg":"stream: finishing up"}
351
+ {"time":"2026-07-23T02:29:02.470200334Z","level":"INFO","msg":"handler: closed"}
352
+ {"time":"2026-07-23T02:29:02.470273871Z","level":"INFO","msg":"sender: closed"}
353
+ {"time":"2026-07-23T02:29:02.470277743Z","level":"INFO","msg":"stream: all finished"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_013342-pgr3eari/logs/debug.log CHANGED
@@ -18,3 +18,48 @@ config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2'
18
  2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2442] Wrapping output streams.
19
  2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2465] Redirects installed.
20
  2026-07-23 01:33:43,706 INFO MainThread:814719 [wandb_init.py:init():1010] run started, returning control to user process
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
  2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2442] Wrapping output streams.
19
  2026-07-23 01:33:43,705 INFO MainThread:814719 [wandb_run.py:_redirect():2465] Redirects installed.
20
  2026-07-23 01:33:43,706 INFO MainThread:814719 [wandb_init.py:init():1010] run started, returning control to user process
21
+ 2026-07-23 02:28:50,770 INFO wandb-AsyncioManager-main:814719 [service_client.py:_forward_responses():122] Reached EOF.
22
+ 2026-07-23 02:28:50,770 INFO wandb-AsyncioManager-main:814719 [mailbox.py:close():154] Closing mailbox, abandoning 2 handles.
23
+ 2026-07-23 02:28:50,770 ERROR wandb-AsyncioManager-main:814719 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
24
+ Traceback (most recent call last):
25
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
26
+ await fn()
27
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 90, in loop
28
+ await asyncio_compat.race(
29
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 278, in race
30
+ async with open_task_group(race=True) as tg:
31
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^
32
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/contextlib.py", line 217, in __aexit__
33
+ await anext(self.gen)
34
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 240, in open_task_group
35
+ await task_group._wait_all(race=race, timeout=exit_timeout)
36
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 180, in _wait_all
37
+ raise exc
38
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 113, in _print_all
39
+ result = await handle.wait_async(timeout=None)
40
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
41
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox_handle.py", line 127, in wait_async
42
+ response = await self._handle.wait_async(timeout=timeout)
43
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
44
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/response_handle.py", line 123, in wait_async
45
+ raise HandleAbandonedError()
46
+ wandb.sdk.mailbox.mailbox_handle.HandleAbandonedError
47
+ 2026-07-23 02:28:59,234 ERROR wandb-AsyncioManager-main:814719 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
48
+ Traceback (most recent call last):
49
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
50
+ await fn()
51
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 87, in loop
52
+ should_stop = await self._check_should_stop()
53
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
54
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 115, in _check_should_stop
55
+ handle = await self._interface.deliver_async(request)
56
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
57
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/interface/interface_sock.py", line 58, in deliver_async
58
+ handle = await self._client.deliver(request)
59
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
60
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 57, in deliver
61
+ handle = self._mailbox.require_response(request)
62
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
63
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox.py", line 89, in require_response
64
+ raise MailboxClosedError()
65
+ wandb.sdk.mailbox.mailbox.MailboxClosedError
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/files/wandb-summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"_wandb":{"runtime":0},"_runtime":0}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug-core.log ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-07-23T02:29:37.218272849Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpbi3491l3/port-2499418.txt","pid":2499418,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
2
+ {"time":"2026-07-23T02:29:37.218902899Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":2499418}
3
+ {"time":"2026-07-23T02:29:37.218864854Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-2499418-2500321-1544228546/socket","Net":"unix"}}
4
+ {"time":"2026-07-23T02:29:37.402297125Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
5
+ {"time":"2026-07-23T02:29:37.409066123Z","level":"INFO","msg":"handleInformInit: received","streamId":"eb0sb786","id":"1(@)"}
6
+ {"time":"2026-07-23T02:29:37.868916316Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"eb0sb786","id":"1(@)"}
7
+ {"time":"2026-07-23T02:29:37.870710243Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpewnffvrr/port-2499551.txt","pid":2499551,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
8
+ {"time":"2026-07-23T02:29:37.871145347Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-2499551-2500408-2757815951/socket","Net":"unix"}}
9
+ {"time":"2026-07-23T02:29:37.871202129Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":2499551}
10
+ {"time":"2026-07-23T02:29:38.054483267Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
11
+ {"time":"2026-07-23T02:29:38.064116376Z","level":"INFO","msg":"handleInformInit: received","streamId":"cemf0rge","id":"1(@)"}
12
+ {"time":"2026-07-23T02:29:38.350897746Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
13
+ {"time":"2026-07-23T02:29:38.351103797Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
14
+ {"time":"2026-07-23T02:29:38.351133399Z","level":"INFO","msg":"server is shutting down"}
15
+ {"time":"2026-07-23T02:29:38.351150646Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
16
+ {"time":"2026-07-23T02:29:38.351193796Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
17
+ {"time":"2026-07-23T02:29:38.351238698Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-2499418-2500321-1544228546/socket","Net":"unix"}}
18
+ {"time":"2026-07-23T02:29:38.352704212Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
19
+ {"time":"2026-07-23T02:29:38.352724918Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
20
+ {"time":"2026-07-23T02:29:38.352730597Z","level":"INFO","msg":"server is closed"}
21
+ {"time":"2026-07-23T02:29:38.445119298Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"cemf0rge","id":"1(@)"}
22
+ {"time":"2026-07-23T02:29:38.796444376Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
23
+ {"time":"2026-07-23T02:29:38.796598651Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
24
+ {"time":"2026-07-23T02:29:38.796618722Z","level":"INFO","msg":"server is shutting down"}
25
+ {"time":"2026-07-23T02:29:38.796624128Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
26
+ {"time":"2026-07-23T02:29:38.796668926Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
27
+ {"time":"2026-07-23T02:29:38.796751175Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-2499551-2500408-2757815951/socket","Net":"unix"}}
28
+ {"time":"2026-07-23T02:29:38.797992882Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
29
+ {"time":"2026-07-23T02:29:38.798016256Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
30
+ {"time":"2026-07-23T02:29:38.798022072Z","level":"INFO","msg":"server is closed"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug-internal.log ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-07-23T02:29:38.064274108Z","level":"INFO","msg":"wandb-core"}
2
+ {"time":"2026-07-23T02:29:38.06450816Z","level":"INFO","msg":"stream: starting","core version":"0.27.2"}
3
+ {"time":"2026-07-23T02:29:38.44495026Z","level":"INFO","msg":"stream: created new stream","id":"cemf0rge"}
4
+ {"time":"2026-07-23T02:29:38.445115186Z","level":"INFO","msg":"stream: started"}
5
+ {"time":"2026-07-23T02:29:38.44526429Z","level":"INFO","msg":"writer: started","stream_id":"cemf0rge"}
6
+ {"time":"2026-07-23T02:29:38.445290711Z","level":"INFO","msg":"handler: started"}
7
+ {"time":"2026-07-23T02:29:38.445315115Z","level":"INFO","msg":"sender: started"}
8
+ {"time":"2026-07-23T02:29:38.786524565Z","level":"ERROR","msg":"runupserter: failed to init run","error":"no data but must resume"}
9
+ {"time":"2026-07-23T02:29:38.796867757Z","level":"ERROR","msg":"sender: uploadConfigFile: runhandle: run not yet initialized"}
10
+ {"time":"2026-07-23T02:29:38.797204677Z","level":"ERROR","msg":"runfiles: runhandle: run not yet initialized"}
11
+ {"time":"2026-07-23T02:29:38.797284772Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
12
+ {"time":"2026-07-23T02:29:38.797292885Z","level":"INFO","msg":"stream: finishing up"}
13
+ {"time":"2026-07-23T02:29:38.797301918Z","level":"INFO","msg":"handler: closed"}
14
+ {"time":"2026-07-23T02:29:38.797325603Z","level":"INFO","msg":"sender: closed"}
15
+ {"time":"2026-07-23T02:29:38.797328171Z","level":"INFO","msg":"stream: all finished"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug.log ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_setup.py:_flush():81] Current SDK version is 0.27.2
2
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_setup.py:_flush():81] Configure stats pid to 2499551
3
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug.log
5
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/logs/debug-internal.log
6
+ 2026-07-23 02:29:37,836 INFO MainThread:2499551 [wandb_init.py:init():766] calling init triggers
7
+ 2026-07-23 02:29:37,837 INFO MainThread:2499551 [wandb_init.py:init():771] wandb.init called with sweep_config: {}
8
+ config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2', 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2', 'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 1.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 1.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'torchcodec', 'return_uint8': False, 'depth_output_unit': 'mm', 'streaming': False, 'eval_split': 0.0}, 'env': None, 'policy': {'type': 'groot_rkd', 'n_obs_steps': 1, 'input_features': {'observation.state': {'type': <FeatureType.STATE: 'STATE'>, 'shape': [16]}, 'observation.images.wrist': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.right_shoulder': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.guide': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}}, 'output_features': {'action': {'type': <FeatureType.ACTION: 'ACTION'>, 'shape': [10]}}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': None, 'private': None, 'tags': None, 'license': None, 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model', 'pretrained_revision': None, 'chunk_size': 16, 'n_action_steps': 16, 'max_state_dim': 132, 'max_action_dim': 132, 'normalization_mapping': {'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>}, 'base_model_path': 'nvidia/GR00T-N1.7-3B', 'action_decode_transform': None, 'embodiment_tag': 'new_embodiment', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'tune_vlln': True, 'tune_top_llm_layers': 0, 'num_inference_timesteps': None, 'rtc_ramp_rate': None, 'use_flash_attention': False, 'use_relative_actions': False, 'relative_exclude_joints': [], 'optimizer_lr': 0.0001, 'optimizer_betas': [0.9, 0.999], 'optimizer_eps': 1e-08, 'optimizer_weight_decay': 1e-05, 'warmup_steps': 3000, 'warmup_ratio': 0.05, 'use_bf16': True, 'model_params_fp32': True, 'image_size': [256, 256], 'tokenizer_assets_repo': None, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_full_model': False, 'video_backend': 'decord', 'balance_dataset_weights': True, 'balance_trajectory_weights': True, 'dataset_paths': None, 'output_dir': './tmp/gr00t', 'save_steps': 1000, 'max_steps': 10000, 'batch_size': 32, 'dataloader_num_workers': 8, 'report_to': 'wandb', 'resume': False, 'action_expert_num_layers': 16, 'rkd_enabled': True, 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11', 'rkd_student_pooling': 'attention', 'rkd_student_attention_hidden_dim': 512, 'rkd_student_attention_output_dim': 512, 'rkd_fm_loss_weight': 1.0, 'rkd_loss_weight': 1.0, 'rkd_distance_weight': 1.0, 'rkd_angle_weight': 2.0}, 'reward_model': None, 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K', 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60', 'resume': True, 'seed': 42, 'cudnn_deterministic': False, 'num_workers': 4, 'batch_size': 64, 'prefetch_factor': 4, 'persistent_workers': True, 'steps': 60000, 'env_eval_freq': 20000, 'log_freq': 200, 'eval_steps': 0, 'max_eval_samples': 0, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 20000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 0.0001, 'weight_decay': 1e-05, 'grad_clip_norm': 1.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': {'type': 'diffuser', 'num_warmup_steps': 3000, 'name': 'cosine'}, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': True, 'recording': False, 'recording_repo_id': None, 'recording_private': False}, 'wandb': {'enable': True, 'disable_artifact': True, 'project': 'lerobot', 'entity': None, 'notes': None, 'run_id': 'cemf0rge', 'mode': None, 'add_tags': True}, 'peft': None, 'job': {'target': None, 'image': 'huggingface/lerobot-gpu:latest', 'timeout': '2d', 'detach': False, 'tags': []}, 'save_checkpoint_to_hub': False, 'sample_weighting': None, 'rename_map': {}, 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000', '_wandb': {}}
9
+ 2026-07-23 02:29:37,837 INFO MainThread:2499551 [wandb_init.py:init():814] starting backend
10
+ 2026-07-23 02:29:38,054 INFO MainThread:2499551 [wandb_init.py:init():829] sending inform_init request
11
+ 2026-07-23 02:29:38,446 INFO MainThread:2499551 [wandb_init.py:init():834] backend started and connected
12
+ 2026-07-23 02:29:38,447 INFO MainThread:2499551 [wandb_init.py:init():904] updated telemetry
13
+ 2026-07-23 02:29:38,453 INFO MainThread:2499551 [wandb_init.py:init():927] communicating run to backend with 90.0 second timeout
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_022937-cemf0rge/run-cemf0rge.wandb ADDED
Binary file (5.07 kB). View file
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/output.log ADDED
@@ -0,0 +1,149 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ INFO 2026-07-23 02:31:26 db_utils.py:126 Logs will be synced with wandb.
2
+ INFO 2026-07-23 02:31:26 db_utils.py:127 Track this run --> https://wandb.ai/minje227_hyu-hanyang-university/lerobot/runs/rs1egwv8
3
+ INFO 2026-07-23 02:31:26 ot_train.py:307 Creating dataset
4
+ INFO 2026-07-23 02:31:29 ot_train.py:341 Creating policy
5
+ INFO 2026-07-23 02:31:29 ng_groot.py:193 The Groot policy wraps NVIDIA's GR00T n1.7 model. Loading pretrained model from: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model
6
+ INFO 2026-07-23 02:31:29 ng_groot.py:227 Detected fine-tuned LeRobot checkpoint, loading with state dict...
7
+ Fetching 27 files: 100%|██████████| 27/27 [00:00<00:00, 1256.77it/s]
8
+ `torch_dtype` is deprecated! Use `dtype` instead!
9
+ Loading weights: 100%|██████████| 807/807 [00:01<00:00, 580.05it/s]
10
+ GR00TN17 LOAD REPORT from: /home/ext_minje/.cache/huggingface/hub/models--nvidia--GR00T-N1.7-3B/snapshots/2fc962b973bccdd5d8ce4f67cc63b264d6886495
11
+ Key | Status | |
12
+ ---------------------------------------------------------------------+------------+--+-
13
+ action_head.model.transformer_blocks.{16...31}.attn1.to_out.0.bias | UNEXPECTED | |
14
+ action_head.model.transformer_blocks.{16...31}.ff.net.2.bias | UNEXPECTED | |
15
+ action_head.model.transformer_blocks.{16...31}.attn1.to_q.weight | UNEXPECTED | |
16
+ action_head.model.transformer_blocks.{16...31}.ff.net.0.proj.weight | UNEXPECTED | |
17
+ action_head.model.transformer_blocks.{16...31}.norm1.linear.bias | UNEXPECTED | |
18
+ action_head.model.transformer_blocks.{16...31}.attn1.to_q.bias | UNEXPECTED | |
19
+ action_head.model.transformer_blocks.{16...31}.attn1.to_k.bias | UNEXPECTED | |
20
+ action_head.model.transformer_blocks.{16...31}.attn1.to_v.weight | UNEXPECTED | |
21
+ action_head.model.transformer_blocks.{16...31}.attn1.to_out.0.weight | UNEXPECTED | |
22
+ action_head.model.transformer_blocks.{16...31}.attn1.to_v.bias | UNEXPECTED | |
23
+ action_head.model.transformer_blocks.{16...31}.norm1.linear.weight | UNEXPECTED | |
24
+ action_head.model.transformer_blocks.{16...31}.attn1.to_k.weight | UNEXPECTED | |
25
+ action_head.model.transformer_blocks.{16...31}.ff.net.0.proj.bias | UNEXPECTED | |
26
+ action_head.model.transformer_blocks.{16...31}.ff.net.2.weight | UNEXPECTED | |
27
+
28
+ Notes:
29
+ - UNEXPECTED: can be ignored when loading from different task/architecture; not ok if you expect identical arch.
30
+ Loading weights from local directory
31
+ INFO 2026-07-23 02:31:45 ot_train.py:414 Creating optimizer and scheduler
32
+ INFO 2026-07-23 02:31:47 ot_train.py:448 Output dir: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
33
+ INFO 2026-07-23 02:31:47 ot_train.py:455 cfg.steps=60000 (60K)
34
+ INFO 2026-07-23 02:31:47 ot_train.py:456 dataset.num_frames=218367 (218K)
35
+ INFO 2026-07-23 02:31:47 ot_train.py:457 dataset.num_episodes=4930
36
+ INFO 2026-07-23 02:31:47 ot_train.py:460 Effective batch size: 64 x 1 = 64
37
+ INFO 2026-07-23 02:31:47 ot_train.py:461 num_learnable_params=1080755329 (1B)
38
+ INFO 2026-07-23 02:31:47 ot_train.py:462 num_total_params=2604374177 (3B)
39
+ INFO 2026-07-23 02:31:47 ot_train.py:503 Resuming data order at epoch 13, sample 42176
40
+ Training: 0%| | 0/30000 [00:00<?, ?step/s]INFO 2026-07-23 02:31:48 ot_train.py:613 Start offline training on a fixed dataset, with effective batch size: 64
41
+ Training: 1%| | 200/30000 [05:04<11:42:04, 1.41s/step]INFO 2026-07-23 02:36:52 ot_train.py:658 step:30K smpl:2M ep:44K epch:8.85 loss:0.041 grdn:0.319 lr:5.4e-05 updt_s:0.810 data_s:0.712 smp/s:42 mem_gb:29.84
42
+ Training: 1%|▏ | 400/30000 [09:58<11:50:43, 1.44s/step]INFO 2026-07-23 02:41:46 ot_train.py:658 step:30K smpl:2M ep:44K epch:8.91 loss:0.040 grdn:0.308 lr:5.3e-05 updt_s:0.777 data_s:0.689 smp/s:44 mem_gb:29.84
43
+ Training: 2%|▏ | 600/30000 [14:54<11:47:28, 1.44s/step]INFO 2026-07-23 02:46:42 ot_train.py:658 step:31K smpl:2M ep:44K epch:8.97 loss:0.040 grdn:0.310 lr:5.3e-05 updt_s:0.775 data_s:0.698 smp/s:43 mem_gb:29.84
44
+ Training: 3%|▎ | 800/30000 [19:52<12:08:23, 1.50s/step]INFO 2026-07-23 02:51:40 ot_train.py:658 step:31K smpl:2M ep:45K epch:9.03 loss:0.040 grdn:0.301 lr:5.2e-05 updt_s:0.772 data_s:0.716 smp/s:43 mem_gb:29.84
45
+ Training: 3%|▎ | 1000/30000 [26:19<28:06:38, 3.49s/step]INFO 2026-07-23 02:58:07 ot_train.py:658 step:31K smpl:2M ep:45K epch:9.09 loss:0.039 grdn:0.292 lr:5.2e-05 updt_s:0.762 data_s:1.168 smp/s:33 mem_gb:29.84
46
+ Training: 4%|▍ | 1200/30000 [34:20<17:28:09, 2.18s/step]INFO 2026-07-23 03:06:08 ot_train.py:658 step:31K smpl:2M ep:45K epch:9.14 loss:0.040 grdn:0.304 lr:5.1e-05 updt_s:0.786 data_s:1.620 smp/s:27 mem_gb:29.84
47
+ Training: 5%|▍ | 1400/30000 [40:26<11:57:15, 1.50s/step]INFO 2026-07-23 03:12:14 ot_train.py:658 step:31K smpl:2M ep:45K epch:9.20 loss:0.040 grdn:0.336 lr:5.1e-05 updt_s:0.776 data_s:1.048 smp/s:35 mem_gb:29.84
48
+ Training: 5%|▌ | 1600/30000 [46:11<27:09:05, 3.44s/step]INFO 2026-07-23 03:17:59 ot_train.py:658 step:32K smpl:2M ep:46K epch:9.26 loss:0.039 grdn:0.302 lr:5.0e-05 updt_s:0.825 data_s:0.896 smp/s:37 mem_gb:29.82
49
+ Training: 6%|▌ | 1800/30000 [54:38<24:33:23, 3.13s/step]INFO 2026-07-23 03:26:26 ot_train.py:658 step:32K smpl:2M ep:46K epch:9.32 loss:0.037 grdn:0.286 lr:4.9e-05 updt_s:0.823 data_s:1.707 smp/s:25 mem_gb:29.85
50
+ Training: 7%|▋ | 2000/30000 [1:01:43<14:06:08, 1.81s/step]INFO 2026-07-23 03:33:31 ot_train.py:658 step:32K smpl:2M ep:46K epch:9.38 loss:0.037 grdn:0.297 lr:4.9e-05 updt_s:0.807 data_s:1.319 smp/s:30 mem_gb:29.85
51
+ Training: 7%|▋ | 2200/30000 [1:08:07<13:17:12, 1.72s/step]INFO 2026-07-23 03:39:55 ot_train.py:658 step:32K smpl:2M ep:47K epch:9.44 loss:0.037 grdn:0.289 lr:4.8e-05 updt_s:0.828 data_s:1.087 smp/s:33 mem_gb:29.85
52
+ Training: 8%|▊ | 2400/30000 [1:13:30<9:55:12, 1.29s/step]INFO 2026-07-23 03:45:18 ot_train.py:658 step:32K smpl:2M ep:47K epch:9.50 loss:0.037 grdn:0.298 lr:4.8e-05 updt_s:0.809 data_s:0.801 smp/s:40 mem_gb:29.85
53
+ Training: 9%|▊ | 2600/30000 [1:18:14<9:50:52, 1.29s/step]INFO 2026-07-23 03:50:02 ot_train.py:658 step:33K smpl:2M ep:47K epch:9.55 loss:0.037 grdn:0.286 lr:4.7e-05 updt_s:0.862 data_s:0.553 smp/s:45 mem_gb:29.85
54
+ Training: 9%|▉ | 2800/30000 [1:22:35<9:47:49, 1.30s/step]INFO 2026-07-23 03:54:23 ot_train.py:658 step:33K smpl:2M ep:47K epch:9.61 loss:0.036 grdn:0.301 lr:4.7e-05 updt_s:0.765 data_s:0.539 smp/s:49 mem_gb:29.84
55
+ Training: 10%|█ | 3000/30000 [1:27:06<9:32:39, 1.27s/step]INFO 2026-07-23 03:58:54 ot_train.py:658 step:33K smpl:2M ep:48K epch:9.67 loss:0.036 grdn:0.282 lr:4.6e-05 updt_s:0.772 data_s:0.578 smp/s:47 mem_gb:29.85
56
+ Training: 11%|█ | 3200/30000 [1:32:14<33:05:48, 4.45s/step]INFO 2026-07-23 04:04:02 ot_train.py:658 step:33K smpl:2M ep:48K epch:9.73 loss:0.036 grdn:0.287 lr:4.6e-05 updt_s:0.888 data_s:0.648 smp/s:42 mem_gb:29.84
57
+ Training: 11%|█▏ | 3400/30000 [1:38:58<10:00:45, 1.36s/step]INFO 2026-07-23 04:10:46 ot_train.py:658 step:33K smpl:2M ep:48K epch:9.79 loss:0.037 grdn:0.296 lr:4.5e-05 updt_s:1.109 data_s:0.908 smp/s:32 mem_gb:29.85
58
+ Training: 12%|█▏ | 3600/30000 [1:43:42<9:21:23, 1.28s/step]INFO 2026-07-23 04:15:30 ot_train.py:658 step:34K smpl:2M ep:49K epch:9.85 loss:0.035 grdn:0.276 lr:4.4e-05 updt_s:0.799 data_s:0.618 smp/s:45 mem_gb:29.84
59
+ Training: 13%|█▎ | 3800/30000 [1:48:01<9:15:45, 1.27s/step]INFO 2026-07-23 04:19:49 ot_train.py:658 step:34K smpl:2M ep:49K epch:9.91 loss:0.035 grdn:0.291 lr:4.4e-05 updt_s:0.745 data_s:0.546 smp/s:50 mem_gb:29.84
60
+ Training: 13%|█▎ | 4000/30000 [1:52:21<9:11:59, 1.27s/step]INFO 2026-07-23 04:24:09 ot_train.py:658 step:34K smpl:2M ep:49K epch:9.96 loss:0.034 grdn:0.293 lr:4.3e-05 updt_s:0.744 data_s:0.554 smp/s:49 mem_gb:29.83
61
+ Training: 14%|█▍ | 4200/30000 [1:56:41<9:46:57, 1.37s/step] INFO 2026-07-23 04:28:29 ot_train.py:658 step:34K smpl:2M ep:49K epch:10.02 loss:0.033 grdn:0.270 lr:4.3e-05 updt_s:0.748 data_s:0.550 smp/s:49 mem_gb:29.85
62
+ Training: 15%|█▍ | 4400/30000 [2:00:59<9:23:54, 1.32s/step]INFO 2026-07-23 04:32:48 ot_train.py:658 step:34K smpl:2M ep:50K epch:10.08 loss:0.033 grdn:0.283 lr:4.2e-05 updt_s:0.739 data_s:0.549 smp/s:50 mem_gb:29.85
63
+ Training: 15%|█▌ | 4600/30000 [2:05:24<9:16:51, 1.32s/step]INFO 2026-07-23 04:37:12 ot_train.py:658 step:35K smpl:2M ep:50K epch:10.14 loss:0.033 grdn:0.295 lr:4.2e-05 updt_s:0.753 data_s:0.568 smp/s:48 mem_gb:29.85
64
+ Training: 16%|█▌ | 4800/30000 [2:09:46<9:00:30, 1.29s/step]INFO 2026-07-23 04:41:34 ot_train.py:658 step:35K smpl:2M ep:50K epch:10.20 loss:0.034 grdn:0.310 lr:4.1e-05 updt_s:0.742 data_s:0.563 smp/s:49 mem_gb:29.85
65
+ Training: 17%|█▋ | 5000/30000 [2:14:08<9:19:05, 1.34s/step]INFO 2026-07-23 04:45:56 ot_train.py:658 step:35K smpl:2M ep:51K epch:10.26 loss:0.034 grdn:0.291 lr:4.1e-05 updt_s:0.730 data_s:0.578 smp/s:49 mem_gb:29.85
66
+ INFO 2026-07-23 04:45:56 ot_train.py:703 Checkpoint policy after step 35000
67
+ INFO 2026-07-23 04:46:20 in_yaml.py:1413 Saved verified rolling checkpoint step=35000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00
68
+ Training: 17%|█▋ | 5200/30000 [2:18:55<8:50:43, 1.28s/step]INFO 2026-07-23 04:50:43 ot_train.py:658 step:35K smpl:2M ep:51K epch:10.32 loss:0.033 grdn:0.286 lr:4.0e-05 updt_s:0.774 data_s:0.539 smp/s:49 mem_gb:29.84
69
+ Training: 18%|█▊ | 5400/30000 [2:23:14<8:57:44, 1.31s/step]INFO 2026-07-23 04:55:02 ot_train.py:658 step:35K smpl:2M ep:51K epch:10.38 loss:0.033 grdn:0.283 lr:4.0e-05 updt_s:0.775 data_s:0.514 smp/s:50 mem_gb:29.84
70
+ Training: 19%|█▊ | 5600/30000 [2:27:37<8:55:10, 1.32s/step]INFO 2026-07-23 04:59:25 ot_train.py:658 step:36K smpl:2M ep:51K epch:10.43 loss:0.033 grdn:0.290 lr:3.9e-05 updt_s:0.784 data_s:0.530 smp/s:49 mem_gb:29.85
71
+ Training: 19%|█▉ | 5800/30000 [2:32:04<8:54:23, 1.32s/step]INFO 2026-07-23 05:03:52 ot_train.py:658 step:36K smpl:2M ep:52K epch:10.49 loss:0.033 grdn:0.281 lr:3.9e-05 updt_s:0.765 data_s:0.569 smp/s:48 mem_gb:29.85
72
+ Training: 20%|██ | 6000/30000 [2:37:07<10:46:18, 1.62s/step]INFO 2026-07-23 05:08:55 ot_train.py:658 step:36K smpl:2M ep:52K epch:10.55 loss:0.032 grdn:0.288 lr:3.8e-05 updt_s:0.824 data_s:0.687 smp/s:42 mem_gb:29.85
73
+ Training: 21%|██ | 6200/30000 [2:42:34<10:40:44, 1.62s/step]INFO 2026-07-23 05:14:22 ot_train.py:658 step:36K smpl:2M ep:52K epch:10.61 loss:0.031 grdn:0.290 lr:3.7e-05 updt_s:0.841 data_s:0.787 smp/s:39 mem_gb:29.83
74
+ Training: 21%|██▏ | 6400/30000 [2:48:24<10:48:23, 1.65s/step]INFO 2026-07-23 05:20:12 ot_train.py:658 step:36K smpl:2M ep:53K epch:10.67 loss:0.031 grdn:0.277 lr:3.7e-05 updt_s:0.876 data_s:0.868 smp/s:37 mem_gb:29.85
75
+ Training: 22%|██▏ | 6600/30000 [2:53:49<10:25:44, 1.60s/step]INFO 2026-07-23 05:25:37 ot_train.py:658 step:37K smpl:2M ep:53K epch:10.73 loss:0.030 grdn:0.260 lr:3.6e-05 updt_s:0.846 data_s:0.777 smp/s:39 mem_gb:29.85
76
+ Training: 23%|██▎ | 6800/30000 [2:59:31<10:51:40, 1.69s/step]INFO 2026-07-23 05:31:19 ot_train.py:658 step:37K smpl:2M ep:53K epch:10.79 loss:0.030 grdn:0.275 lr:3.6e-05 updt_s:0.862 data_s:0.845 smp/s:38 mem_gb:29.85
77
+ Training: 23%|██▎ | 7000/30000 [3:04:57<10:34:46, 1.66s/step]INFO 2026-07-23 05:36:45 ot_train.py:658 step:37K smpl:2M ep:53K epch:10.84 loss:0.030 grdn:0.276 lr:3.5e-05 updt_s:0.834 data_s:0.792 smp/s:39 mem_gb:29.85
78
+ Training: 24%|██▍ | 7200/30000 [3:10:20<10:05:08, 1.59s/step]INFO 2026-07-23 05:42:08 ot_train.py:658 step:37K smpl:2M ep:54K epch:10.90 loss:0.030 grdn:0.268 lr:3.5e-05 updt_s:0.836 data_s:0.771 smp/s:40 mem_gb:29.85
79
+ Training: 25%|██▍ | 7400/30000 [3:15:40<10:13:24, 1.63s/step]INFO 2026-07-23 05:47:28 ot_train.py:658 step:37K smpl:2M ep:54K epch:10.96 loss:0.030 grdn:0.269 lr:3.4e-05 updt_s:0.834 data_s:0.764 smp/s:40 mem_gb:29.85
80
+ Training: 25%|██▌ | 7600/30000 [3:20:55<9:42:54, 1.56s/step]INFO 2026-07-23 05:52:43 ot_train.py:658 step:38K smpl:2M ep:54K epch:11.02 loss:0.030 grdn:0.268 lr:3.4e-05 updt_s:0.816 data_s:0.758 smp/s:41 mem_gb:29.84
81
+ Training: 26%|██▌ | 7800/30000 [3:26:13<9:42:03, 1.57s/step]INFO 2026-07-23 05:58:01 ot_train.py:658 step:38K smpl:2M ep:55K epch:11.08 loss:0.030 grdn:0.291 lr:3.3e-05 updt_s:0.805 data_s:0.777 smp/s:40 mem_gb:29.85
82
+ Training: 27%|██▋ | 8000/30000 [3:31:27<10:16:26, 1.68s/step]INFO 2026-07-23 06:03:15 ot_train.py:658 step:38K smpl:2M ep:55K epch:11.14 loss:0.030 grdn:0.284 lr:3.3e-05 updt_s:0.789 data_s:0.780 smp/s:41 mem_gb:29.84
83
+ Training: 27%|██▋ | 8200/30000 [3:36:42<9:23:31, 1.55s/step]INFO 2026-07-23 06:08:30 ot_train.py:658 step:38K smpl:2M ep:55K epch:11.20 loss:0.030 grdn:0.284 lr:3.2e-05 updt_s:0.840 data_s:0.729 smp/s:41 mem_gb:29.84
84
+ Training: 28%|██▊ | 8400/30000 [3:41:56<8:53:52, 1.48s/step]INFO 2026-07-23 06:13:44 ot_train.py:658 step:38K smpl:2M ep:55K epch:11.25 loss:0.029 grdn:0.282 lr:3.2e-05 updt_s:0.862 data_s:0.706 smp/s:41 mem_gb:29.83
85
+ Training: 29%|██▊ | 8600/30000 [3:47:16<9:34:08, 1.61s/step]INFO 2026-07-23 06:19:04 ot_train.py:658 step:39K smpl:2M ep:56K epch:11.31 loss:0.028 grdn:0.259 lr:3.1e-05 updt_s:0.822 data_s:0.775 smp/s:40 mem_gb:29.84
86
+ Training: 29%|██▉ | 8800/30000 [3:53:46<27:38:09, 4.69s/step]INFO 2026-07-23 06:25:34 ot_train.py:658 step:39K smpl:2M ep:56K epch:11.37 loss:0.027 grdn:0.265 lr:3.1e-05 updt_s:0.770 data_s:1.176 smp/s:33 mem_gb:29.85
87
+ Training: 30%|███ | 9000/30000 [4:02:29<13:54:12, 2.38s/step]INFO 2026-07-23 06:34:17 ot_train.py:658 step:39K smpl:2M ep:56K epch:11.43 loss:0.027 grdn:0.269 lr:3.0e-05 updt_s:0.786 data_s:1.826 smp/s:24 mem_gb:29.85
88
+ Training: 31%|███ | 9200/30000 [4:09:19<10:05:13, 1.75s/step]INFO 2026-07-23 06:41:07 ot_train.py:658 step:39K smpl:3M ep:57K epch:11.49 loss:0.027 grdn:0.259 lr:3.0e-05 updt_s:0.772 data_s:1.273 smp/s:31 mem_gb:29.85
89
+ Training: 31%|███▏ | 9400/30000 [4:17:43<14:44:45, 2.58s/step]INFO 2026-07-23 06:49:31 ot_train.py:658 step:39K smpl:3M ep:57K epch:11.55 loss:0.027 grdn:0.261 lr:2.9e-05 updt_s:0.739 data_s:1.778 smp/s:25 mem_gb:29.85
90
+ Training: 32%|███▏ | 9600/30000 [4:23:55<9:45:19, 1.72s/step]INFO 2026-07-23 06:55:43 ot_train.py:658 step:40K smpl:3M ep:57K epch:11.61 loss:0.026 grdn:0.270 lr:2.9e-05 updt_s:0.854 data_s:1.001 smp/s:35 mem_gb:29.84
91
+ Training: 33%|███▎ | 9800/30000 [4:29:32<9:19:33, 1.66s/step]INFO 2026-07-23 07:01:20 ot_train.py:658 step:40K smpl:3M ep:58K epch:11.66 loss:0.027 grdn:0.270 lr:2.8e-05 updt_s:0.820 data_s:0.858 smp/s:38 mem_gb:29.84
92
+ Training: 33%|███▎ | 10000/30000 [4:35:41<9:49:56, 1.77s/step]INFO 2026-07-23 07:07:29 ot_train.py:658 step:40K smpl:3M ep:58K epch:11.72 loss:0.026 grdn:0.253 lr:2.8e-05 updt_s:0.834 data_s:1.007 smp/s:35 mem_gb:29.85
93
+ INFO 2026-07-23 07:07:29 ot_train.py:703 Checkpoint policy after step 40000
94
+ INFO 2026-07-23 07:07:58 in_yaml.py:1413 Saved verified rolling checkpoint step=40000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01
95
+ Training: 34%|███▍ | 10200/30000 [4:42:42<8:51:50, 1.61s/step]INFO 2026-07-23 07:14:31 ot_train.py:658 step:40K smpl:3M ep:58K epch:11.78 loss:0.027 grdn:0.264 lr:2.7e-05 updt_s:0.815 data_s:1.142 smp/s:33 mem_gb:29.84
96
+ Training: 35%|███▍ | 10400/30000 [4:48:12<9:06:45, 1.67s/step]INFO 2026-07-23 07:20:00 ot_train.py:658 step:40K smpl:3M ep:58K epch:11.84 loss:0.026 grdn:0.257 lr:2.7e-05 updt_s:0.838 data_s:0.805 smp/s:39 mem_gb:29.84
97
+ Training: 35%|███▌ | 10600/30000 [4:53:30<8:13:24, 1.53s/step]INFO 2026-07-23 07:25:18 ot_train.py:658 step:41K smpl:3M ep:59K epch:11.90 loss:0.027 grdn:0.266 lr:2.6e-05 updt_s:0.792 data_s:0.792 smp/s:40 mem_gb:29.85
98
+ Training: 36%|███▌ | 10800/30000 [4:58:55<9:23:44, 1.76s/step]INFO 2026-07-23 07:30:43 ot_train.py:658 step:41K smpl:3M ep:59K epch:11.96 loss:0.024 grdn:0.242 lr:2.6e-05 updt_s:0.810 data_s:0.815 smp/s:39 mem_gb:29.83
99
+ Training: 37%|███▋ | 11000/30000 [5:04:48<9:02:44, 1.71s/step]INFO 2026-07-23 07:36:36 ot_train.py:658 step:41K smpl:3M ep:59K epch:12.02 loss:0.025 grdn:0.246 lr:2.5e-05 updt_s:0.870 data_s:0.890 smp/s:36 mem_gb:29.85
100
+ Training: 37%|███▋ | 11200/30000 [5:10:37<8:47:55, 1.68s/step]INFO 2026-07-23 07:42:25 ot_train.py:658 step:41K smpl:3M ep:60K epch:12.08 loss:0.025 grdn:0.249 lr:2.5e-05 updt_s:0.850 data_s:0.891 smp/s:37 mem_gb:29.84
101
+ Training: 38%|███▊ | 11400/30000 [5:16:48<18:00:35, 3.49s/step]INFO 2026-07-23 07:48:36 ot_train.py:658 step:41K smpl:3M ep:60K epch:12.13 loss:0.024 grdn:0.246 lr:2.4e-05 updt_s:0.818 data_s:1.029 smp/s:35 mem_gb:29.85
102
+ Training: 39%|███▊ | 11600/30000 [5:24:19<11:00:04, 2.15s/step]INFO 2026-07-23 07:56:07 ot_train.py:658 step:42K smpl:3M ep:60K epch:12.19 loss:0.024 grdn:0.258 lr:2.4e-05 updt_s:0.789 data_s:1.466 smp/s:28 mem_gb:29.85
103
+ Training: 39%|███▉ | 11800/30000 [5:30:49<16:03:56, 3.18s/step]INFO 2026-07-23 08:02:37 ot_train.py:658 step:42K smpl:3M ep:60K epch:12.25 loss:0.024 grdn:0.246 lr:2.3e-05 updt_s:0.792 data_s:1.154 smp/s:33 mem_gb:29.85
104
+ Training: 40%|████ | 12000/30000 [5:37:13<8:10:00, 1.63s/step]INFO 2026-07-23 08:09:01 ot_train.py:658 step:42K smpl:3M ep:61K epch:12.31 loss:0.024 grdn:0.243 lr:2.3e-05 updt_s:0.796 data_s:1.117 smp/s:33 mem_gb:29.85
105
+ Training: 41%|████ | 12200/30000 [5:43:13<14:44:41, 2.98s/step]INFO 2026-07-23 08:15:01 ot_train.py:658 step:42K smpl:3M ep:61K epch:12.37 loss:0.023 grdn:0.255 lr:2.2e-05 updt_s:0.802 data_s:0.994 smp/s:36 mem_gb:29.85
106
+ Training: 41%|████▏ | 12400/30000 [5:48:53<9:33:34, 1.96s/step] INFO 2026-07-23 08:20:41 ot_train.py:658 step:42K smpl:3M ep:61K epch:12.43 loss:0.024 grdn:0.250 lr:2.2e-05 updt_s:0.832 data_s:0.865 smp/s:38 mem_gb:29.84
107
+ Training: 42%|████▏ | 12600/30000 [5:56:03<13:31:05, 2.80s/step]INFO 2026-07-23 08:27:51 ot_train.py:658 step:43K smpl:3M ep:62K epch:12.49 loss:0.025 grdn:0.248 lr:2.2e-05 updt_s:0.807 data_s:1.341 smp/s:30 mem_gb:29.84
108
+ Training: 43%|████▎ | 12800/30000 [6:06:50<60:52:50, 12.74s/step]INFO 2026-07-23 08:38:38 ot_train.py:658 step:43K smpl:3M ep:62K epch:12.54 loss:0.024 grdn:0.252 lr:2.1e-05 updt_s:0.798 data_s:2.432 smp/s:20 mem_gb:29.84
109
+ Training: 43%|████▎ | 13000/30000 [6:14:13<7:49:04, 1.66s/step]INFO 2026-07-23 08:46:01 ot_train.py:658 step:43K smpl:3M ep:62K epch:12.60 loss:0.023 grdn:0.256 lr:2.1e-05 updt_s:0.821 data_s:1.391 smp/s:29 mem_gb:29.83
110
+ Training: 44%|████▍ | 13200/30000 [6:19:43<7:34:42, 1.62s/step]INFO 2026-07-23 08:51:31 ot_train.py:658 step:43K smpl:3M ep:62K epch:12.66 loss:0.021 grdn:0.227 lr:2.0e-05 updt_s:0.812 data_s:0.834 smp/s:39 mem_gb:29.85
111
+ Training: 45%|████▍ | 13400/30000 [6:25:08<7:51:44, 1.71s/step]INFO 2026-07-23 08:56:56 ot_train.py:658 step:43K smpl:3M ep:63K epch:12.72 loss:0.022 grdn:0.244 lr:2.0e-05 updt_s:0.804 data_s:0.816 smp/s:39 mem_gb:29.85
112
+ Training: 45%|████▌ | 13600/30000 [6:30:32<7:26:31, 1.63s/step]INFO 2026-07-23 09:02:20 ot_train.py:658 step:44K smpl:3M ep:63K epch:12.78 loss:0.022 grdn:0.225 lr:1.9e-05 updt_s:0.802 data_s:0.813 smp/s:40 mem_gb:29.85
113
+ Training: 46%|████▌ | 13800/30000 [6:36:38<7:49:33, 1.74s/step]INFO 2026-07-23 09:08:26 ot_train.py:658 step:44K smpl:3M ep:63K epch:12.84 loss:0.022 grdn:0.260 lr:1.9e-05 updt_s:0.789 data_s:1.036 smp/s:35 mem_gb:29.84
114
+ Training: 47%|████▋ | 14000/30000 [6:42:39<7:57:32, 1.79s/step]INFO 2026-07-23 09:14:27 ot_train.py:658 step:44K smpl:3M ep:64K epch:12.90 loss:0.022 grdn:0.231 lr:1.8e-05 updt_s:0.811 data_s:0.993 smp/s:35 mem_gb:29.84
115
+ Training: 47%|████▋ | 14200/30000 [6:48:23<8:18:40, 1.89s/step]INFO 2026-07-23 09:20:11 ot_train.py:658 step:44K smpl:3M ep:64K epch:12.95 loss:0.022 grdn:0.246 lr:1.8e-05 updt_s:0.830 data_s:0.883 smp/s:37 mem_gb:29.84
116
+ Training: 48%|████▊ | 14400/30000 [6:55:47<9:01:51, 2.08s/step]INFO 2026-07-23 09:27:35 ot_train.py:658 step:44K smpl:3M ep:64K epch:13.01 loss:0.021 grdn:0.230 lr:1.8e-05 updt_s:0.799 data_s:1.417 smp/s:29 mem_gb:29.85
117
+ Training: 49%|████▊ | 14600/30000 [7:01:25<6:27:45, 1.51s/step]INFO 2026-07-23 09:33:13 ot_train.py:658 step:45K smpl:3M ep:64K epch:13.07 loss:0.021 grdn:0.236 lr:1.7e-05 updt_s:0.804 data_s:0.885 smp/s:38 mem_gb:29.85
118
+ Training: 49%|████▉ | 14800/30000 [7:06:44<7:06:24, 1.68s/step]INFO 2026-07-23 09:38:32 ot_train.py:658 step:45K smpl:3M ep:65K epch:13.13 loss:0.021 grdn:0.246 lr:1.7e-05 updt_s:0.824 data_s:0.765 smp/s:40 mem_gb:29.85
119
+ Training: 50%|█████ | 15000/30000 [7:12:11<6:49:49, 1.64s/step]INFO 2026-07-23 09:43:59 ot_train.py:658 step:45K smpl:3M ep:65K epch:13.19 loss:0.022 grdn:0.227 lr:1.6e-05 updt_s:0.829 data_s:0.802 smp/s:39 mem_gb:29.85
120
+ INFO 2026-07-23 09:43:59 ot_train.py:703 Checkpoint policy after step 45000
121
+ INFO 2026-07-23 09:44:25 in_yaml.py:1413 Saved verified rolling checkpoint step=45000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00
122
+ Training: 51%|█████ | 15200/30000 [7:18:00<7:01:12, 1.71s/step]INFO 2026-07-23 09:49:48 ot_train.py:658 step:45K smpl:3M ep:65K epch:13.25 loss:0.021 grdn:0.233 lr:1.6e-05 updt_s:0.812 data_s:0.800 smp/s:40 mem_gb:29.83
123
+ Training: 51%|█████▏ | 15400/30000 [7:23:32<6:44:54, 1.66s/step]INFO 2026-07-23 09:55:20 ot_train.py:658 step:45K smpl:3M ep:66K epch:13.31 loss:0.020 grdn:0.218 lr:1.6e-05 updt_s:0.859 data_s:0.793 smp/s:39 mem_gb:29.85
124
+ Training: 52%|█████▏ | 15600/30000 [7:28:51<6:04:24, 1.52s/step]INFO 2026-07-23 10:00:39 ot_train.py:658 step:46K smpl:3M ep:66K epch:13.36 loss:0.020 grdn:0.226 lr:1.5e-05 updt_s:0.814 data_s:0.777 smp/s:40 mem_gb:29.85
125
+ Training: 53%|█████▎ | 15800/30000 [7:34:16<6:18:25, 1.60s/step]INFO 2026-07-23 10:06:04 ot_train.py:658 step:46K smpl:3M ep:66K epch:13.42 loss:0.020 grdn:0.241 lr:1.5e-05 updt_s:0.816 data_s:0.806 smp/s:39 mem_gb:29.85
126
+ Training: 53%|█████▎ | 16000/30000 [7:40:32<6:54:11, 1.78s/step]INFO 2026-07-23 10:12:20 ot_train.py:658 step:46K smpl:3M ep:66K epch:13.48 loss:0.020 grdn:0.223 lr:1.4e-05 updt_s:0.801 data_s:1.074 smp/s:34 mem_gb:29.84
127
+ Training: 54%|█████▍ | 16200/30000 [7:47:00<6:49:38, 1.78s/step]INFO 2026-07-23 10:18:48 ot_train.py:658 step:46K smpl:3M ep:67K epch:13.54 loss:0.020 grdn:0.233 lr:1.4e-05 updt_s:0.857 data_s:1.081 smp/s:33 mem_gb:29.85
128
+ Training: 55%|█████▍ | 16400/30000 [7:53:59<6:23:57, 1.69s/step]INFO 2026-07-23 10:25:47 ot_train.py:658 step:46K smpl:3M ep:67K epch:13.60 loss:0.020 grdn:0.228 lr:1.4e-05 updt_s:0.825 data_s:1.264 smp/s:31 mem_gb:29.85
129
+ Training: 55%|█████▌ | 16600/30000 [8:01:35<6:11:47, 1.66s/step]INFO 2026-07-23 10:33:23 ot_train.py:658 step:47K smpl:3M ep:67K epch:13.66 loss:0.019 grdn:0.227 lr:1.3e-05 updt_s:0.838 data_s:1.439 smp/s:28 mem_gb:29.85
130
+ Training: 56%|█████▌ | 16800/30000 [8:07:39<5:53:47, 1.61s/step]INFO 2026-07-23 10:39:27 ot_train.py:658 step:47K smpl:3M ep:68K epch:13.72 loss:0.020 grdn:0.228 lr:1.3e-05 updt_s:0.856 data_s:0.960 smp/s:35 mem_gb:29.84
131
+ Training: 57%|█████▋ | 17000/30000 [8:12:56<6:00:32, 1.66s/step]INFO 2026-07-23 10:44:44 ot_train.py:658 step:47K smpl:3M ep:68K epch:13.77 loss:0.019 grdn:0.228 lr:1.2e-05 updt_s:0.889 data_s:0.692 smp/s:40 mem_gb:29.84
132
+ Training: 57%|█████▋ | 17200/30000 [8:27:00<17:27:42, 4.91s/step]INFO 2026-07-23 10:58:48 ot_train.py:658 step:47K smpl:3M ep:68K epch:13.83 loss:0.020 grdn:0.228 lr:1.2e-05 updt_s:3.367 data_s:0.850 smp/s:15 mem_gb:29.85
133
+ Training: 58%|█████▊ | 17400/30000 [8:41:30<13:58:08, 3.99s/step]INFO 2026-07-23 11:13:18 ot_train.py:658 step:47K smpl:3M ep:68K epch:13.89 loss:0.019 grdn:0.225 lr:1.2e-05 updt_s:3.355 data_s:0.992 smp/s:15 mem_gb:29.83
134
+ Training: 59%|█████▊ | 17600/30000 [8:55:45<18:19:28, 5.32s/step]INFO 2026-07-23 11:27:33 ot_train.py:658 step:48K smpl:3M ep:69K epch:13.95 loss:0.019 grdn:0.236 lr:1.1e-05 updt_s:3.422 data_s:0.849 smp/s:15 mem_gb:29.85
135
+ Training: 59%|█████▉ | 17800/30000 [9:11:33<13:40:01, 4.03s/step]INFO 2026-07-23 11:43:21 ot_train.py:658 step:48K smpl:3M ep:69K epch:14.01 loss:0.018 grdn:0.225 lr:1.1e-05 updt_s:3.145 data_s:1.591 smp/s:14 mem_gb:29.85
136
+ Training: 60%|██████ | 18000/30000 [9:25:29<13:42:28, 4.11s/step]INFO 2026-07-23 11:57:17 ot_train.py:658 step:48K smpl:3M ep:69K epch:14.07 loss:0.019 grdn:0.227 lr:1.1e-05 updt_s:3.320 data_s:0.854 smp/s:15 mem_gb:29.84
137
+ Training: 61%|██████ | 18200/30000 [9:39:49<14:15:11, 4.35s/step]INFO 2026-07-23 12:11:37 ot_train.py:658 step:48K smpl:3M ep:70K epch:14.13 loss:0.018 grdn:0.230 lr:1.0e-05 updt_s:3.414 data_s:0.886 smp/s:15 mem_gb:29.85
138
+ Training: 61%|██████▏ | 18400/30000 [9:53:56<14:28:13, 4.49s/step]INFO 2026-07-23 12:25:44 ot_train.py:658 step:48K smpl:3M ep:70K epch:14.19 loss:0.018 grdn:0.218 lr:1.0e-05 updt_s:3.465 data_s:0.766 smp/s:15 mem_gb:29.85
139
+ Training: 62%|██████▏ | 18600/30000 [10:07:35<13:01:01, 4.11s/step]INFO 2026-07-23 12:39:23 ot_train.py:658 step:49K smpl:3M ep:70K epch:14.24 loss:0.018 grdn:0.226 lr:9.7e-06 updt_s:3.424 data_s:0.669 smp/s:16 mem_gb:29.85
140
+ Training: 63%|██████▎ | 18800/30000 [10:21:17<13:50:28, 4.45s/step]INFO 2026-07-23 12:53:05 ot_train.py:658 step:49K smpl:3M ep:71K epch:14.30 loss:0.018 grdn:0.217 lr:9.4e-06 updt_s:3.456 data_s:0.649 smp/s:16 mem_gb:29.84
141
+ Training: 63%|██████▎ | 19000/30000 [10:35:21<13:52:13, 4.54s/step]INFO 2026-07-23 13:07:09 ot_train.py:658 step:49K smpl:3M ep:71K epch:14.36 loss:0.018 grdn:0.209 lr:9.1e-06 updt_s:3.570 data_s:0.644 smp/s:15 mem_gb:29.84
142
+ Training: 64%|██████▍ | 19200/30000 [10:48:56<10:50:18, 3.61s/step]INFO 2026-07-23 13:20:44 ot_train.py:658 step:49K smpl:3M ep:71K epch:14.42 loss:0.018 grdn:0.219 lr:8.8e-06 updt_s:3.473 data_s:0.600 smp/s:16 mem_gb:29.84
143
+ Training: 65%|██████▍ | 19400/30000 [11:03:04<12:04:34, 4.10s/step]INFO 2026-07-23 13:34:52 ot_train.py:658 step:49K smpl:3M ep:71K epch:14.48 loss:0.018 grdn:0.217 lr:8.4e-06 updt_s:3.473 data_s:0.765 smp/s:15 mem_gb:29.84
144
+ Training: 65%|██████▌ | 19600/30000 [11:17:00<10:49:27, 3.75s/step]INFO 2026-07-23 13:48:48 ot_train.py:658 step:50K smpl:3M ep:72K epch:14.54 loss:0.017 grdn:0.226 lr:8.1e-06 updt_s:3.486 data_s:0.694 smp/s:15 mem_gb:29.85
145
+ Training: 66%|██████▌ | 19800/30000 [11:30:48<12:12:01, 4.31s/step]INFO 2026-07-23 14:02:36 ot_train.py:658 step:50K smpl:3M ep:72K epch:14.60 loss:0.017 grdn:0.216 lr:7.8e-06 updt_s:3.490 data_s:0.643 smp/s:15 mem_gb:29.83
146
+ Training: 67%|██████▋ | 20000/30000 [11:44:24<11:03:22, 3.98s/step]INFO 2026-07-23 14:16:12 ot_train.py:658 step:50K smpl:3M ep:72K epch:14.65 loss:0.017 grdn:0.210 lr:7.5e-06 updt_s:3.495 data_s:0.583 smp/s:16 mem_gb:29.84
147
+ INFO 2026-07-23 14:16:12 ot_train.py:703 Checkpoint policy after step 50000
148
+ INFO 2026-07-23 14:16:43 in_yaml.py:1413 Saved verified rolling checkpoint step=50000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01
149
+ Training: 67%|██████▋ | 20082/30000 [11:50:29<10:53:43, 3.95s/step]
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/requirements.txt ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ tzdata==2026.3
2
+ nvidia-cufile-cu12==1.13.1.3
3
+ wandb==0.27.2
4
+ six==1.17.0
5
+ peft==0.19.1
6
+ nvidia-curand-cu12==10.3.9.90
7
+ nvidia-cuda-runtime-cu12==12.8.90
8
+ datasets==4.8.5
9
+ attrs==26.1.0
10
+ cloudpickle==3.1.2
11
+ packaging==25.0
12
+ nvidia-cublas-cu12==12.8.4.1
13
+ cmake==4.1.3
14
+ psutil==7.2.2
15
+ pydantic==2.13.4
16
+ GitPython==3.1.51
17
+ pillow==12.2.0
18
+ pytz==2026.2
19
+ cuda-pathfinder==1.2.2
20
+ nvidia-cusparse-cu12==12.5.8.93
21
+ typing_extensions==4.15.0
22
+ urllib3==2.7.0
23
+ pydantic_core==2.46.4
24
+ draccus==0.10.0
25
+ mypy_extensions==1.1.0
26
+ dill==0.4.1
27
+ typer==0.26.8
28
+ decord==0.6.0
29
+ tqdm==4.68.4
30
+ triton==3.6.0
31
+ idna==3.18
32
+ requests==2.34.2
33
+ nvidia-cudnn-cu12==9.19.0.56
34
+ mdurl==0.1.2
35
+ shellingham==1.5.4
36
+ zipp==4.1.0
37
+ timm==1.0.28
38
+ torch==2.11.0+cu128
39
+ tokenizers==0.22.2
40
+ nvidia-cusparselt-cu12==0.7.1
41
+ nvidia-cufft-cu12==11.3.3.83
42
+ nvidia-cusolver-cu12==11.7.3.90
43
+ opencv-python-headless==4.13.0.92
44
+ httpx==0.28.1
45
+ Jinja2==3.1.6
46
+ einops==0.8.2
47
+ mergedeep==1.3.4
48
+ platformdirs==4.10.0
49
+ nvidia-nvjitlink-cu12==12.8.93
50
+ lerobot==0.6.0
51
+ PyYAML==6.0.3
52
+ accelerate==1.14.0
53
+ anyio==4.14.2
54
+ gymnasium==1.3.0
55
+ pip==26.1.2
56
+ huggingface_hub==1.23.0
57
+ diffusers==0.35.2
58
+ annotated-doc==0.0.4
59
+ pyarrow==25.0.0
60
+ python-dateutil==2.9.0.post0
61
+ termcolor==3.3.0
62
+ nvidia-nccl-cu12==2.28.9
63
+ sentry-sdk==2.65.0
64
+ torchcodec==0.11.1
65
+ pyyaml-include==1.4.1
66
+ multiprocess==0.70.19
67
+ wrapt==2.2.2
68
+ typing-inspection==0.4.2
69
+ nvidia-cuda-nvrtc-cu12==12.8.93
70
+ Pygments==2.20.0
71
+ nvidia-cuda-cupti-cu12==12.8.90
72
+ certifi==2026.6.17
73
+ setuptools==78.1.0
74
+ charset-normalizer==3.4.9
75
+ propcache==0.5.2
76
+ multidict==6.7.1
77
+ regex==2026.7.10
78
+ frozenlist==1.8.0
79
+ smmap==5.0.3
80
+ hf-xet==1.5.1
81
+ torchvision==0.26.0+cu128
82
+ gitdb==4.0.12
83
+ transformers==5.5.4
84
+ rich==15.0.0
85
+ cuda-bindings==12.9.4
86
+ protobuf==7.35.1
87
+ aiosignal==1.4.0
88
+ jsonlines==4.0.0
89
+ yarl==1.24.2
90
+ filelock==3.29.0
91
+ av==15.1.0
92
+ networkx==3.6.1
93
+ sympy==1.14.0
94
+ wheel==0.47.0
95
+ nvidia-nvtx-cu12==12.8.90
96
+ xxhash==3.8.1
97
+ toml==0.10.2
98
+ MarkupSafe==3.0.3
99
+ pandas==2.3.3
100
+ safetensors==0.8.0
101
+ click==8.4.2
102
+ fsspec==2026.2.0
103
+ dm-tree==0.1.10
104
+ absl-py==2.5.0
105
+ Farama-Notifications==0.0.6
106
+ mpmath==1.3.0
107
+ importlib_metadata==9.0.0
108
+ markdown-it-py==4.2.0
109
+ h11==0.16.0
110
+ typing-inspect==0.9.0
111
+ httpcore==1.0.9
112
+ annotated-types==0.7.0
113
+ numpy==2.2.6
114
+ nvidia-nvshmem-cu12==3.4.5
115
+ aiohappyeyeballs==2.7.1
116
+ cuda-toolkit==12.8.1
117
+ aiohttp==3.14.1
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/files/wandb-metadata.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-6.8.0-55-generic-x86_64-with-glibc2.39",
3
+ "python": "CPython 3.12.13",
4
+ "startedAt": "2026-07-23T02:31:25.059506Z",
5
+ "args": [
6
+ "--config_path=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model/train_config.json",
7
+ "--resume=true",
8
+ "--output_dir=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K",
9
+ "--dataset.root=/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2",
10
+ "--wandb.run_id=rs1egwv8"
11
+ ],
12
+ "program": "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py",
13
+ "codePath": "benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py",
14
+ "codePathLocal": "my_scripts/lerobot_train_yaml.py",
15
+ "git": {
16
+ "remote": "git@github.com:paragon7060/clvla.git",
17
+ "commit": "a329728c2c9524d79b966c4a7a82ee56a4c2abf7"
18
+ },
19
+ "root": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K",
20
+ "host": "ext_minje-box",
21
+ "executable": "/home/ext_minje/miniconda3/envs/lerobot060_groot/bin/python",
22
+ "cpu_count": 112,
23
+ "cpu_count_logical": 224,
24
+ "gpu": "NVIDIA H200",
25
+ "gpu_count": 2,
26
+ "disk": {
27
+ "/": {
28
+ "total": "1888425144320",
29
+ "used": "1138103234560"
30
+ }
31
+ },
32
+ "memory": {
33
+ "total": "2164176568320"
34
+ },
35
+ "gpu_nvidia": [
36
+ {
37
+ "name": "NVIDIA H200",
38
+ "memoryTotal": "150754820096",
39
+ "cudaCores": 16896,
40
+ "architecture": "Hopper",
41
+ "uuid": "GPU-86be6e5b-fb48-fa1c-9bd7-83d336cb2991"
42
+ },
43
+ {
44
+ "name": "NVIDIA H200",
45
+ "memoryTotal": "150754820096",
46
+ "cudaCores": 16896,
47
+ "architecture": "Hopper",
48
+ "uuid": "GPU-24eb1ba2-7313-ff93-1435-748768ebdbdc"
49
+ }
50
+ ],
51
+ "cudaVersion": "12.8",
52
+ "writerId": "rcam6hzew8b6nlqutd1m9r4avsa0csph"
53
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug-core.log ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-07-23T02:31:25.095933268Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmphmwi8v6n/port-2501080.txt","pid":2501080,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
2
+ {"time":"2026-07-23T02:31:25.096932569Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":2501080}
3
+ {"time":"2026-07-23T02:31:25.096927661Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-2501080-2501865-1162647384/socket","Net":"unix"}}
4
+ {"time":"2026-07-23T02:31:25.169461847Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmp4t2rwfks/port-2500816.txt","pid":2500816,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
5
+ {"time":"2026-07-23T02:31:25.170134264Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-2500816-2501888-288489369/socket","Net":"unix"}}
6
+ {"time":"2026-07-23T02:31:25.170259549Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":2500816}
7
+ {"time":"2026-07-23T02:31:25.27857788Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
8
+ {"time":"2026-07-23T02:31:25.285691576Z","level":"INFO","msg":"handleInformInit: received","streamId":"rs1egwv8","id":"1(@)"}
9
+ {"time":"2026-07-23T02:31:25.348726979Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
10
+ {"time":"2026-07-23T02:31:25.356108831Z","level":"INFO","msg":"handleInformInit: received","streamId":"x6da8gup","id":"1(@)"}
11
+ {"time":"2026-07-23T02:31:25.661339067Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"rs1egwv8","id":"1(@)"}
12
+ {"time":"2026-07-23T02:31:25.811527394Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"x6da8gup","id":"1(@)"}
13
+ {"time":"2026-07-23T02:31:31.694097963Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"7skxzehhhgck"}
14
+ {"time":"2026-07-23T02:31:31.708133393Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"b3q6giuvspld"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug-internal.log ADDED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug.log ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_setup.py:_flush():81] Current SDK version is 0.27.2
2
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_setup.py:_flush():81] Configure stats pid to 2501080
3
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug.log
5
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_023125-rs1egwv8/logs/debug-internal.log
6
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_init.py:init():766] calling init triggers
7
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_init.py:init():771] wandb.init called with sweep_config: {}
8
+ config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2', 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2', 'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 1.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 1.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'torchcodec', 'return_uint8': False, 'depth_output_unit': 'mm', 'streaming': False, 'eval_split': 0.0}, 'env': None, 'policy': {'type': 'groot_rkd', 'n_obs_steps': 1, 'input_features': {'observation.state': {'type': <FeatureType.STATE: 'STATE'>, 'shape': [16]}, 'observation.images.wrist': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.right_shoulder': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.guide': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}}, 'output_features': {'action': {'type': <FeatureType.ACTION: 'ACTION'>, 'shape': [10]}}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': None, 'private': None, 'tags': None, 'license': None, 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000/pretrained_model', 'pretrained_revision': None, 'chunk_size': 16, 'n_action_steps': 16, 'max_state_dim': 132, 'max_action_dim': 132, 'normalization_mapping': {'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>}, 'base_model_path': 'nvidia/GR00T-N1.7-3B', 'action_decode_transform': None, 'embodiment_tag': 'new_embodiment', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'tune_vlln': True, 'tune_top_llm_layers': 0, 'num_inference_timesteps': None, 'rtc_ramp_rate': None, 'use_flash_attention': False, 'use_relative_actions': False, 'relative_exclude_joints': [], 'optimizer_lr': 0.0001, 'optimizer_betas': [0.9, 0.999], 'optimizer_eps': 1e-08, 'optimizer_weight_decay': 1e-05, 'warmup_steps': 3000, 'warmup_ratio': 0.05, 'use_bf16': True, 'model_params_fp32': True, 'image_size': [256, 256], 'tokenizer_assets_repo': None, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_full_model': False, 'video_backend': 'decord', 'balance_dataset_weights': True, 'balance_trajectory_weights': True, 'dataset_paths': None, 'output_dir': './tmp/gr00t', 'save_steps': 1000, 'max_steps': 10000, 'batch_size': 32, 'dataloader_num_workers': 8, 'report_to': 'wandb', 'resume': False, 'action_expert_num_layers': 16, 'rkd_enabled': True, 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11', 'rkd_student_pooling': 'attention', 'rkd_student_attention_hidden_dim': 512, 'rkd_student_attention_output_dim': 512, 'rkd_fm_loss_weight': 1.0, 'rkd_loss_weight': 1.0, 'rkd_distance_weight': 1.0, 'rkd_angle_weight': 2.0}, 'reward_model': None, 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K', 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60', 'resume': True, 'seed': 42, 'cudnn_deterministic': False, 'num_workers': 4, 'batch_size': 64, 'prefetch_factor': 4, 'persistent_workers': True, 'steps': 60000, 'env_eval_freq': 20000, 'log_freq': 200, 'eval_steps': 0, 'max_eval_samples': 0, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 20000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 0.0001, 'weight_decay': 1e-05, 'grad_clip_norm': 1.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': {'type': 'diffuser', 'num_warmup_steps': 3000, 'name': 'cosine'}, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': True, 'recording': False, 'recording_repo_id': None, 'recording_private': False}, 'wandb': {'enable': True, 'disable_artifact': True, 'project': 'lerobot', 'entity': None, 'notes': None, 'run_id': 'rs1egwv8', 'mode': None, 'add_tags': True}, 'peft': None, 'job': {'target': None, 'image': 'huggingface/lerobot-gpu:latest', 'timeout': '2d', 'detach': False, 'tags': []}, 'save_checkpoint_to_hub': False, 'sample_weighting': None, 'rename_map': {}, 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/030000', '_wandb': {}}
9
+ 2026-07-23 02:31:25,060 INFO MainThread:2501080 [wandb_init.py:init():814] starting backend
10
+ 2026-07-23 02:31:25,278 INFO MainThread:2501080 [wandb_init.py:init():829] sending inform_init request
11
+ 2026-07-23 02:31:25,662 INFO MainThread:2501080 [wandb_init.py:init():834] backend started and connected
12
+ 2026-07-23 02:31:25,664 INFO MainThread:2501080 [wandb_init.py:init():904] updated telemetry
13
+ 2026-07-23 02:31:25,668 INFO MainThread:2501080 [wandb_init.py:init():927] communicating run to backend with 90.0 second timeout
14
+ 2026-07-23 02:31:26,617 INFO MainThread:2501080 [wandb_init.py:init():972] starting run threads in backend
15
+ 2026-07-23 02:31:26,676 INFO MainThread:2501080 [wandb_run.py:_console_start():2523] atexit reg
16
+ 2026-07-23 02:31:26,677 INFO MainThread:2501080 [wandb_run.py:_redirect():2373] redirect: wrap_raw
17
+ 2026-07-23 02:31:26,677 INFO MainThread:2501080 [wandb_run.py:_redirect():2442] Wrapping output streams.
18
+ 2026-07-23 02:31:26,677 INFO MainThread:2501080 [wandb_run.py:_redirect():2465] Redirects installed.
19
+ 2026-07-23 02:31:26,678 INFO MainThread:2501080 [wandb_init.py:init():1010] run started, returning control to user process
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/config.yaml ADDED
@@ -0,0 +1,329 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ _wandb:
2
+ value:
3
+ cli_version: 0.27.2
4
+ e:
5
+ 3vx7sedbdsz3uay6fuuuh2e3qlqtbnr9:
6
+ args:
7
+ - --config_path=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/train_config.json
8
+ - --resume=true
9
+ codePath: clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py
10
+ codePathLocal: clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py
11
+ cpu_count: 112
12
+ cpu_count_logical: 224
13
+ cudaVersion: "12.9"
14
+ disk:
15
+ /:
16
+ total: "1888425144320"
17
+ used: "1327086579712"
18
+ executable: /home/ext_minje/miniconda3/envs/lerobot060_groot/bin/python
19
+ gpu: NVIDIA H200
20
+ gpu_count: 2
21
+ gpu_nvidia:
22
+ - architecture: Hopper
23
+ cudaCores: 16896
24
+ memoryTotal: "150754820096"
25
+ name: NVIDIA H200
26
+ uuid: GPU-86be6e5b-fb48-fa1c-9bd7-83d336cb2991
27
+ - architecture: Hopper
28
+ cudaCores: 16896
29
+ memoryTotal: "150754820096"
30
+ name: NVIDIA H200
31
+ uuid: GPU-24eb1ba2-7313-ff93-1435-748768ebdbdc
32
+ host: ext_minje-box
33
+ memory:
34
+ total: "2164176568320"
35
+ os: Linux-6.8.0-55-generic-x86_64-with-glibc2.39
36
+ program: /home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py
37
+ python: CPython 3.12.13
38
+ root: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
39
+ startedAt: "2026-07-23T16:35:01.611693Z"
40
+ writerId: 3vx7sedbdsz3uay6fuuuh2e3qlqtbnr9
41
+ m: []
42
+ python_version: 3.12.13
43
+ t:
44
+ "1":
45
+ - 1
46
+ - 11
47
+ - 41
48
+ - 49
49
+ - 51
50
+ - 71
51
+ - 83
52
+ - 98
53
+ "2":
54
+ - 1
55
+ - 11
56
+ - 41
57
+ - 49
58
+ - 51
59
+ - 71
60
+ - 83
61
+ - 98
62
+ "3":
63
+ - 5
64
+ - 13
65
+ - 14
66
+ - 15
67
+ - 16
68
+ - 61
69
+ - 62
70
+ "4": 3.12.13
71
+ "5": 0.27.2
72
+ "6": 5.5.4
73
+ "10":
74
+ - 21
75
+ "12": 0.27.2
76
+ "13": linux-x86_64
77
+ batch_size:
78
+ value: 64
79
+ checkpoint_path:
80
+ value: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01
81
+ cudnn_deterministic:
82
+ value: false
83
+ dataset:
84
+ value:
85
+ depth_output_unit: mm
86
+ episodes: null
87
+ eval_split: 0
88
+ image_transforms:
89
+ enable: false
90
+ max_num_transforms: 3
91
+ random_order: false
92
+ tfs:
93
+ affine:
94
+ kwargs:
95
+ degrees:
96
+ - -5
97
+ - 5
98
+ translate:
99
+ - 0.05
100
+ - 0.05
101
+ type: RandomAffine
102
+ weight: 1
103
+ brightness:
104
+ kwargs:
105
+ brightness:
106
+ - 0.8
107
+ - 1.2
108
+ type: ColorJitter
109
+ weight: 1
110
+ contrast:
111
+ kwargs:
112
+ contrast:
113
+ - 0.8
114
+ - 1.2
115
+ type: ColorJitter
116
+ weight: 1
117
+ hue:
118
+ kwargs:
119
+ hue:
120
+ - -0.05
121
+ - 0.05
122
+ type: ColorJitter
123
+ weight: 1
124
+ saturation:
125
+ kwargs:
126
+ saturation:
127
+ - 0.5
128
+ - 1.5
129
+ type: ColorJitter
130
+ weight: 1
131
+ sharpness:
132
+ kwargs:
133
+ sharpness:
134
+ - 0.5
135
+ - 1.5
136
+ type: SharpnessJitter
137
+ weight: 1
138
+ repo_id: Whalswp/INSIGHTfixposV4_filtered_multispace_v2
139
+ return_uint8: false
140
+ revision: null
141
+ root: /home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2
142
+ streaming: false
143
+ use_imagenet_stats: true
144
+ video_backend: torchcodec
145
+ env:
146
+ value: null
147
+ env_eval_freq:
148
+ value: 20000
149
+ eval:
150
+ value:
151
+ batch_size: 50
152
+ n_episodes: 50
153
+ recording: false
154
+ recording_private: false
155
+ recording_repo_id: null
156
+ use_async_envs: true
157
+ eval_steps:
158
+ value: 0
159
+ job:
160
+ value:
161
+ detach: false
162
+ image: huggingface/lerobot-gpu:latest
163
+ tags: []
164
+ target: null
165
+ timeout: 2d
166
+ job_name:
167
+ value: INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60
168
+ log_freq:
169
+ value: 200
170
+ max_eval_samples:
171
+ value: 0
172
+ num_workers:
173
+ value: 4
174
+ optimizer:
175
+ value:
176
+ betas:
177
+ - 0.9
178
+ - 0.999
179
+ eps: 1e-08
180
+ grad_clip_norm: 1
181
+ lr: 0.0001
182
+ type: adamw
183
+ weight_decay: 1e-05
184
+ output_dir:
185
+ value: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
186
+ peft:
187
+ value: null
188
+ persistent_workers:
189
+ value: true
190
+ policy:
191
+ value:
192
+ action_decode_transform: null
193
+ action_expert_num_layers: 16
194
+ balance_dataset_weights: true
195
+ balance_trajectory_weights: true
196
+ base_model_path: nvidia/GR00T-N1.7-3B
197
+ batch_size: 32
198
+ chunk_size: 16
199
+ dataloader_num_workers: 8
200
+ dataset_paths: null
201
+ device: cuda
202
+ embodiment_tag: new_embodiment
203
+ image_size:
204
+ - 256
205
+ - 256
206
+ input_features:
207
+ observation.images.guide:
208
+ shape:
209
+ - 3
210
+ - 224
211
+ - 224
212
+ type: VISUAL
213
+ observation.images.right_shoulder:
214
+ shape:
215
+ - 3
216
+ - 224
217
+ - 224
218
+ type: VISUAL
219
+ observation.images.wrist:
220
+ shape:
221
+ - 3
222
+ - 224
223
+ - 224
224
+ type: VISUAL
225
+ observation.state:
226
+ shape:
227
+ - 16
228
+ type: STATE
229
+ license: null
230
+ lora_alpha: 16
231
+ lora_dropout: 0.1
232
+ lora_full_model: false
233
+ lora_rank: 0
234
+ max_action_dim: 132
235
+ max_state_dim: 132
236
+ max_steps: 10000
237
+ model_params_fp32: true
238
+ n_action_steps: 16
239
+ n_obs_steps: 1
240
+ normalization_mapping:
241
+ ACTION: IDENTITY
242
+ STATE: IDENTITY
243
+ VISUAL: IDENTITY
244
+ num_inference_timesteps: null
245
+ optimizer_betas:
246
+ - 0.9
247
+ - 0.999
248
+ optimizer_eps: 1e-08
249
+ optimizer_lr: 0.0001
250
+ optimizer_weight_decay: 1e-05
251
+ output_dir: ./tmp/gr00t
252
+ output_features:
253
+ action:
254
+ shape:
255
+ - 10
256
+ type: ACTION
257
+ pretrained_path: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model
258
+ pretrained_revision: null
259
+ private: null
260
+ push_to_hub: false
261
+ relative_exclude_joints: []
262
+ repo_id: null
263
+ report_to: wandb
264
+ resume: false
265
+ rkd_angle_weight: 2
266
+ rkd_distance_weight: 1
267
+ rkd_enabled: true
268
+ rkd_fm_loss_weight: 1
269
+ rkd_loss_weight: 1
270
+ rkd_student_attention_hidden_dim: 512
271
+ rkd_student_attention_output_dim: 512
272
+ rkd_student_pooling: attention
273
+ rkd_vae_checkpoint: /home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11
274
+ rtc_ramp_rate: null
275
+ save_steps: 1000
276
+ tags: null
277
+ tokenizer_assets_repo: null
278
+ tune_diffusion_model: true
279
+ tune_llm: false
280
+ tune_projector: true
281
+ tune_top_llm_layers: 0
282
+ tune_visual: false
283
+ tune_vlln: true
284
+ type: groot_rkd
285
+ use_amp: false
286
+ use_bf16: true
287
+ use_flash_attention: false
288
+ use_peft: false
289
+ use_relative_actions: false
290
+ video_backend: decord
291
+ warmup_ratio: 0.05
292
+ warmup_steps: 3000
293
+ prefetch_factor:
294
+ value: 4
295
+ resume:
296
+ value: true
297
+ reward_model:
298
+ value: null
299
+ sample_weighting:
300
+ value: null
301
+ save_checkpoint:
302
+ value: true
303
+ save_checkpoint_to_hub:
304
+ value: false
305
+ save_freq:
306
+ value: 20000
307
+ scheduler:
308
+ value:
309
+ name: cosine
310
+ num_warmup_steps: 3000
311
+ type: diffuser
312
+ seed:
313
+ value: 42
314
+ steps:
315
+ value: 60000
316
+ tolerance_s:
317
+ value: 0.0001
318
+ use_policy_training_preset:
319
+ value: true
320
+ wandb:
321
+ value:
322
+ add_tags: true
323
+ disable_artifact: true
324
+ enable: true
325
+ entity: null
326
+ mode: null
327
+ notes: null
328
+ project: lerobot
329
+ run_id: rs1egwv8
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/output.log ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ INFO 2026-07-23 16:35:03 db_utils.py:126 Logs will be synced with wandb.
2
+ INFO 2026-07-23 16:35:03 db_utils.py:127 Track this run --> https://wandb.ai/minje227_hyu-hanyang-university/lerobot/runs/rs1egwv8
3
+ INFO 2026-07-23 16:35:03 ot_train.py:310 Creating dataset
4
+ INFO 2026-07-23 16:35:05 ot_train.py:344 Creating policy
5
+ INFO 2026-07-23 16:35:05 ng_groot.py:193 The Groot policy wraps NVIDIA's GR00T n1.7 model. Loading pretrained model from: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model
6
+ INFO 2026-07-23 16:35:05 ng_groot.py:227 Detected fine-tuned LeRobot checkpoint, loading with state dict...
7
+ Fetching 27 files: 100%|██████████| 27/27 [00:00<00:00, 5334.76it/s]
8
+ `torch_dtype` is deprecated! Use `dtype` instead!
9
+ Loading weights: 100%|██████████| 807/807 [00:04<00:00, 187.16it/s]
10
+ GR00TN17 LOAD REPORT from: /home/ext_minje/.cache/huggingface/hub/models--nvidia--GR00T-N1.7-3B/snapshots/2fc962b973bccdd5d8ce4f67cc63b264d6886495
11
+ Key | Status | |
12
+ ---------------------------------------------------------------------+------------+--+-
13
+ action_head.model.transformer_blocks.{16...31}.attn1.to_q.bias | UNEXPECTED | |
14
+ action_head.model.transformer_blocks.{16...31}.attn1.to_q.weight | UNEXPECTED | |
15
+ action_head.model.transformer_blocks.{16...31}.attn1.to_out.0.weight | UNEXPECTED | |
16
+ action_head.model.transformer_blocks.{16...31}.attn1.to_out.0.bias | UNEXPECTED | |
17
+ action_head.model.transformer_blocks.{16...31}.ff.net.2.weight | UNEXPECTED | |
18
+ action_head.model.transformer_blocks.{16...31}.attn1.to_k.bias | UNEXPECTED | |
19
+ action_head.model.transformer_blocks.{16...31}.attn1.to_v.weight | UNEXPECTED | |
20
+ action_head.model.transformer_blocks.{16...31}.attn1.to_k.weight | UNEXPECTED | |
21
+ action_head.model.transformer_blocks.{16...31}.norm1.linear.weight | UNEXPECTED | |
22
+ action_head.model.transformer_blocks.{16...31}.ff.net.2.bias | UNEXPECTED | |
23
+ action_head.model.transformer_blocks.{16...31}.ff.net.0.proj.weight | UNEXPECTED | |
24
+ action_head.model.transformer_blocks.{16...31}.norm1.linear.bias | UNEXPECTED | |
25
+ action_head.model.transformer_blocks.{16...31}.attn1.to_v.bias | UNEXPECTED | |
26
+ action_head.model.transformer_blocks.{16...31}.ff.net.0.proj.bias | UNEXPECTED | |
27
+
28
+ Notes:
29
+ - UNEXPECTED: can be ignored when loading from different task/architecture; not ok if you expect identical arch.
30
+ Loading weights from local directory
31
+ INFO 2026-07-23 16:35:27 ot_train.py:417 Creating optimizer and scheduler
32
+ INFO 2026-07-23 16:35:31 ot_train.py:451 Output dir: /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K
33
+ INFO 2026-07-23 16:35:31 ot_train.py:458 cfg.steps=60000 (60K)
34
+ INFO 2026-07-23 16:35:31 ot_train.py:459 dataset.num_frames=218367 (218K)
35
+ INFO 2026-07-23 16:35:31 ot_train.py:460 dataset.num_episodes=4930
36
+ INFO 2026-07-23 16:35:31 ot_train.py:463 Effective batch size: 64 x 1 = 64
37
+ INFO 2026-07-23 16:35:31 ot_train.py:464 num_learnable_params=1080755329 (1B)
38
+ INFO 2026-07-23 16:35:31 ot_train.py:465 num_total_params=2604374177 (3B)
39
+ INFO 2026-07-23 16:35:31 ot_train.py:506 Resuming data order at epoch 22, sample 22144
40
+ Training: 0%| | 0/10000 [00:00<?, ?step/s]INFO 2026-07-23 16:35:31 ot_train.py:616 Start offline training on a fixed dataset, with effective batch size: 64
41
+ Training: 2%|▏ | 200/10000 [13:50<10:59:57, 4.04s/step]INFO 2026-07-23 16:49:22 ot_train.py:661 step:50K smpl:3M ep:73K epch:14.71 loss:0.017 grdn:0.208 lr:7.3e-06 updt_s:3.597 data_s:0.554 smp/s:15 mem_gb:29.84 flow_matching_loss:0.017 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.017 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.066 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.389 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
42
+ Training: 4%|▍ | 400/10000 [27:29<10:27:16, 3.92s/step]INFO 2026-07-23 17:03:00 ot_train.py:661 step:50K smpl:3M ep:73K epch:14.77 loss:0.017 grdn:0.211 lr:7.0e-06 updt_s:3.559 data_s:0.531 smp/s:16 mem_gb:29.84 flow_matching_loss:0.017 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.017 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.063 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.330 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
43
+ Training: 6%|▌ | 600/10000 [41:05<10:16:11, 3.93s/step]INFO 2026-07-23 17:16:36 ot_train.py:661 step:51K smpl:3M ep:73K epch:14.83 loss:0.017 grdn:0.226 lr:6.7e-06 updt_s:3.547 data_s:0.530 smp/s:16 mem_gb:29.84 flow_matching_loss:0.017 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.017 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.061 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.348 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
44
+ Training: 8%|▊ | 800/10000 [54:43<10:41:44, 4.19s/step]INFO 2026-07-23 17:30:15 ot_train.py:661 step:51K smpl:3M ep:73K epch:14.89 loss:0.016 grdn:0.214 lr:6.4e-06 updt_s:3.563 data_s:0.529 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.058 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.331 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
45
+ Training: 10%|█ | 1000/10000 [1:08:20<10:21:49, 4.15s/step]INFO 2026-07-23 17:43:51 ot_train.py:661 step:51K smpl:3M ep:74K epch:14.95 loss:0.016 grdn:0.210 lr:6.2e-06 updt_s:3.548 data_s:0.530 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.056 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.340 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
46
+ Training: 12%|█▏ | 1200/10000 [1:21:55<10:12:47, 4.18s/step]INFO 2026-07-23 17:57:27 ot_train.py:661 step:51K smpl:3M ep:74K epch:15.01 loss:0.016 grdn:0.210 lr:5.9e-06 updt_s:3.543 data_s:0.531 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.053 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.291 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
47
+ Training: 14%|█▍ | 1400/10000 [1:35:34<10:08:33, 4.25s/step]INFO 2026-07-23 18:11:05 ot_train.py:661 step:51K smpl:3M ep:74K epch:15.06 loss:0.016 grdn:0.213 lr:5.6e-06 updt_s:3.560 data_s:0.530 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.051 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.284 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
48
+ Training: 16%|█▌ | 1600/10000 [1:49:11<9:27:12, 4.05s/step]INFO 2026-07-23 18:24:42 ot_train.py:661 step:52K smpl:3M ep:75K epch:15.12 loss:0.016 grdn:0.212 lr:5.4e-06 updt_s:3.564 data_s:0.521 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.049 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.236 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
49
+ Training: 18%|█▊ | 1800/10000 [2:02:49<8:58:00, 3.94s/step]INFO 2026-07-23 18:38:20 ot_train.py:661 step:52K smpl:3M ep:75K epch:15.18 loss:0.017 grdn:0.215 lr:5.1e-06 updt_s:3.563 data_s:0.525 smp/s:16 mem_gb:29.83 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.046 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.253 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
50
+ Training: 20%|██ | 2000/10000 [2:16:31<9:08:44, 4.12s/step]INFO 2026-07-23 18:52:03 ot_train.py:661 step:52K smpl:3M ep:75K epch:15.24 loss:0.016 grdn:0.220 lr:4.9e-06 updt_s:3.584 data_s:0.527 smp/s:16 mem_gb:29.82 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.044 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.249 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
51
+ Training: 22%|██▏ | 2200/10000 [2:30:16<9:17:43, 4.29s/step]INFO 2026-07-23 19:05:48 ot_train.py:661 step:52K smpl:3M ep:75K epch:15.30 loss:0.016 grdn:0.216 lr:4.7e-06 updt_s:3.590 data_s:0.531 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.042 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.216 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
52
+ Training: 24%|██▍ | 2400/10000 [2:43:59<9:00:55, 4.27s/step]INFO 2026-07-23 19:19:31 ot_train.py:661 step:52K smpl:3M ep:76K epch:15.36 loss:0.016 grdn:0.205 lr:4.4e-06 updt_s:3.586 data_s:0.528 smp/s:16 mem_gb:29.85 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.040 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.197 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
53
+ Training: 26%|██▌ | 2600/10000 [2:57:34<8:36:33, 4.19s/step]INFO 2026-07-23 19:33:06 ot_train.py:661 step:53K smpl:3M ep:76K epch:15.42 loss:0.015 grdn:0.201 lr:4.2e-06 updt_s:3.546 data_s:0.527 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.038 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.223 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
54
+ Training: 28%|██▊ | 2800/10000 [3:11:19<8:27:43, 4.23s/step]INFO 2026-07-23 19:46:51 ot_train.py:661 step:53K smpl:3M ep:76K epch:15.47 loss:0.016 grdn:0.206 lr:4.0e-06 updt_s:3.593 data_s:0.530 smp/s:16 mem_gb:29.84 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.036 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.215 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
55
+ Training: 30%|███ | 3000/10000 [3:24:53<8:18:47, 4.28s/step]INFO 2026-07-23 20:00:24 ot_train.py:661 step:53K smpl:3M ep:77K epch:15.53 loss:0.016 grdn:0.211 lr:3.8e-06 updt_s:3.536 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.034 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.220 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
56
+ Training: 32%|███▏ | 3200/10000 [3:38:26<7:18:43, 3.87s/step]INFO 2026-07-23 20:13:57 ot_train.py:661 step:53K smpl:3M ep:77K epch:15.59 loss:0.016 grdn:0.219 lr:3.6e-06 updt_s:3.536 data_s:0.528 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.032 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.196 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
57
+ Training: 34%|███▍ | 3400/10000 [3:52:03<7:36:05, 4.15s/step]INFO 2026-07-23 20:27:35 ot_train.py:661 step:53K smpl:3M ep:77K epch:15.65 loss:0.016 grdn:0.205 lr:3.4e-06 updt_s:3.557 data_s:0.527 smp/s:16 mem_gb:29.85 flow_matching_loss:0.016 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.016 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.030 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.206 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
58
+ Training: 36%|███▌ | 3600/10000 [4:05:39<7:04:13, 3.98s/step]INFO 2026-07-23 20:41:11 ot_train.py:661 step:54K smpl:3M ep:77K epch:15.71 loss:0.015 grdn:0.208 lr:3.2e-06 updt_s:3.551 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.029 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.184 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
59
+ Training: 38%|███▊ | 3800/10000 [4:19:18<6:57:11, 4.04s/step]INFO 2026-07-23 20:54:50 ot_train.py:661 step:54K smpl:3M ep:78K epch:15.77 loss:0.015 grdn:0.198 lr:3.0e-06 updt_s:3.566 data_s:0.526 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.027 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.188 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
60
+ Training: 40%|████ | 4000/10000 [4:32:56<6:54:03, 4.14s/step]INFO 2026-07-23 21:08:28 ot_train.py:661 step:54K smpl:3M ep:78K epch:15.83 loss:0.016 grdn:0.206 lr:2.8e-06 updt_s:3.556 data_s:0.533 smp/s:16 mem_gb:29.84 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.025 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.214 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
61
+ Training: 42%|████▏ | 4200/10000 [4:46:35<6:24:18, 3.98s/step]INFO 2026-07-23 21:22:07 ot_train.py:661 step:54K smpl:3M ep:78K epch:15.89 loss:0.015 grdn:0.207 lr:2.6e-06 updt_s:3.561 data_s:0.531 smp/s:16 mem_gb:29.83 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.024 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.198 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
62
+ Training: 44%|████▍ | 4400/10000 [5:00:15<6:04:48, 3.91s/step]INFO 2026-07-23 21:35:47 ot_train.py:661 step:54K smpl:3M ep:79K epch:15.94 loss:0.015 grdn:0.204 lr:2.4e-06 updt_s:3.570 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.022 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.194 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
63
+ Training: 46%|████▌ | 4600/10000 [5:13:56<6:00:53, 4.01s/step]INFO 2026-07-23 21:49:28 ot_train.py:661 step:55K smpl:3M ep:79K epch:16.00 loss:0.015 grdn:0.206 lr:2.3e-06 updt_s:3.574 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.021 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.198 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
64
+ Training: 48%|████▊ | 4800/10000 [5:27:34<6:00:41, 4.16s/step]INFO 2026-07-23 22:03:06 ot_train.py:661 step:55K smpl:4M ep:79K epch:16.06 loss:0.015 grdn:0.207 lr:2.1e-06 updt_s:3.560 data_s:0.527 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.019 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.200 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
65
+ Training: 50%|█████ | 5000/10000 [5:41:10<4:26:09, 3.19s/step]INFO 2026-07-23 22:16:42 ot_train.py:661 step:55K smpl:4M ep:79K epch:16.12 loss:0.015 grdn:0.206 lr:2.0e-06 updt_s:3.550 data_s:0.530 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.018 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.203 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
66
+ INFO 2026-07-23 22:16:42 ot_train.py:707 Checkpoint policy after step 55000
67
+ INFO 2026-07-23 22:17:00 in_yaml.py:1414 Saved verified rolling checkpoint step=55000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_00
68
+ Training: 52%|█████▏ | 5200/10000 [5:55:11<5:27:42, 4.10s/step]INFO 2026-07-23 22:30:43 ot_train.py:661 step:55K smpl:4M ep:80K epch:16.18 loss:0.015 grdn:0.200 lr:1.8e-06 updt_s:3.582 data_s:0.530 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.016 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.189 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
69
+ Training: 54%|█████▍ | 5400/10000 [6:08:50<5:03:22, 3.96s/step]INFO 2026-07-23 22:44:21 ot_train.py:661 step:55K smpl:4M ep:80K epch:16.24 loss:0.015 grdn:0.215 lr:1.7e-06 updt_s:3.561 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.015 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.212 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
70
+ Training: 56%|█████▌ | 5600/10000 [6:22:31<5:13:40, 4.28s/step]INFO 2026-07-23 22:58:02 ot_train.py:661 step:56K smpl:4M ep:80K epch:16.30 loss:0.015 grdn:0.206 lr:1.5e-06 updt_s:3.573 data_s:0.530 smp/s:16 mem_gb:29.84 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.014 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.192 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
71
+ Training: 58%|█████▊ | 5800/10000 [6:36:07<4:38:04, 3.97s/step]INFO 2026-07-23 23:11:39 ot_train.py:661 step:56K smpl:4M ep:81K epch:16.35 loss:0.015 grdn:0.205 lr:1.4e-06 updt_s:3.551 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.013 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.155 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
72
+ Training: 60%|██████ | 6000/10000 [6:49:37<4:33:48, 4.11s/step]INFO 2026-07-23 23:25:09 ot_train.py:661 step:56K smpl:4M ep:81K epch:16.41 loss:0.015 grdn:0.201 lr:1.3e-06 updt_s:3.520 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.011 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.122 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
73
+ Training: 62%|██████▏ | 6200/10000 [7:03:15<4:17:22, 4.06s/step]INFO 2026-07-23 23:38:47 ot_train.py:661 step:56K smpl:4M ep:81K epch:16.47 loss:0.015 grdn:0.201 lr:1.2e-06 updt_s:3.556 data_s:0.530 smp/s:16 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.010 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.126 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
74
+ Training: 64%|██████▍ | 6400/10000 [7:16:53<4:03:36, 4.06s/step]INFO 2026-07-23 23:52:24 ot_train.py:661 step:56K smpl:4M ep:81K epch:16.53 loss:0.015 grdn:0.199 lr:1.0e-06 updt_s:3.558 data_s:0.527 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.009 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.127 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
75
+ Training: 66%|██████▌ | 6600/10000 [7:30:32<3:53:04, 4.11s/step]INFO 2026-07-24 00:06:03 ot_train.py:661 step:57K smpl:4M ep:82K epch:16.59 loss:0.015 grdn:0.200 lr:9.3e-07 updt_s:3.565 data_s:0.529 smp/s:16 mem_gb:29.83 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.008 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.126 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
76
+ Training: 68%|██████▊ | 6800/10000 [7:44:09<3:46:36, 4.25s/step]INFO 2026-07-24 00:19:40 ot_train.py:661 step:57K smpl:4M ep:82K epch:16.65 loss:0.015 grdn:0.207 lr:8.2e-07 updt_s:3.551 data_s:0.531 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.007 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.137 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
77
+ Training: 70%|███████ | 7000/10000 [7:57:46<3:25:39, 4.11s/step]INFO 2026-07-24 00:33:17 ot_train.py:661 step:57K smpl:4M ep:82K epch:16.71 loss:0.015 grdn:0.207 lr:7.3e-07 updt_s:3.554 data_s:0.529 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.007 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.128 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
78
+ Training: 72%|███████▏ | 7200/10000 [8:11:20<3:15:16, 4.18s/step]INFO 2026-07-24 00:46:51 ot_train.py:661 step:57K smpl:4M ep:83K epch:16.76 loss:0.015 grdn:0.201 lr:6.4e-07 updt_s:3.535 data_s:0.532 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.006 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.146 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
79
+ Training: 74%|███████▍ | 7400/10000 [8:24:58<2:50:34, 3.94s/step]INFO 2026-07-24 01:00:30 ot_train.py:661 step:57K smpl:4M ep:83K epch:16.82 loss:0.015 grdn:0.201 lr:5.5e-07 updt_s:3.552 data_s:0.539 smp/s:16 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.005 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.119 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
80
+ Training: 76%|███████▌ | 7600/10000 [8:38:55<2:39:14, 3.98s/step]INFO 2026-07-24 01:14:27 ot_train.py:661 step:58K smpl:4M ep:83K epch:16.88 loss:0.014 grdn:0.205 lr:4.7e-07 updt_s:3.595 data_s:0.586 smp/s:15 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.004 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.128 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
81
+ Training: 78%|███████▊ | 7800/10000 [8:52:32<2:24:37, 3.94s/step]INFO 2026-07-24 01:28:03 ot_train.py:661 step:58K smpl:4M ep:84K epch:16.94 loss:0.014 grdn:0.199 lr:4.0e-07 updt_s:3.538 data_s:0.544 smp/s:16 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.004 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.134 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
82
+ Training: 80%|████████ | 8000/10000 [9:06:27<2:13:40, 4.01s/step]INFO 2026-07-24 01:41:59 ot_train.py:661 step:58K smpl:4M ep:84K epch:17.00 loss:0.015 grdn:0.209 lr:3.3e-07 updt_s:3.560 data_s:0.612 smp/s:15 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.003 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.131 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
83
+ Training: 82%|████████▏ | 8200/10000 [9:20:26<2:00:16, 4.01s/step]INFO 2026-07-24 01:55:57 ot_train.py:661 step:58K smpl:4M ep:84K epch:17.06 loss:0.015 grdn:0.202 lr:2.7e-07 updt_s:3.582 data_s:0.609 smp/s:15 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.002 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.130 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
84
+ Training: 84%|████████▍ | 8400/10000 [9:36:41<2:06:49, 4.76s/step]INFO 2026-07-24 02:12:12 ot_train.py:661 step:58K smpl:4M ep:84K epch:17.12 loss:0.014 grdn:0.193 lr:2.2e-07 updt_s:3.983 data_s:0.887 smp/s:13 mem_gb:29.84 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.002 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.127 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
85
+ Training: 86%|████████▌ | 8600/10000 [9:51:48<1:37:09, 4.16s/step]INFO 2026-07-24 02:27:20 ot_train.py:661 step:59K smpl:4M ep:85K epch:17.17 loss:0.015 grdn:0.212 lr:1.7e-07 updt_s:3.602 data_s:0.930 smp/s:14 mem_gb:29.84 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.002 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.149 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
86
+ Training: 88%|████████▊ | 8800/10000 [10:05:55<1:21:37, 4.08s/step]INFO 2026-07-24 02:41:27 ot_train.py:661 step:59K smpl:4M ep:85K epch:17.23 loss:0.015 grdn:0.202 lr:1.3e-07 updt_s:3.527 data_s:0.705 smp/s:15 mem_gb:29.83 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.001 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.132 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
87
+ Training: 90%|█████████ | 9000/10000 [10:20:01<1:15:04, 4.50s/step]INFO 2026-07-24 02:55:32 ot_train.py:661 step:59K smpl:4M ep:85K epch:17.29 loss:0.014 grdn:0.197 lr:9.2e-08 updt_s:3.561 data_s:0.662 smp/s:15 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.001 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.136 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
88
+ Training: 92%|█████████▏| 9200/10000 [10:34:22<59:02, 4.43s/step]INFO 2026-07-24 03:09:54 ot_train.py:661 step:59K smpl:4M ep:86K epch:17.35 loss:0.014 grdn:0.190 lr:6.2e-08 updt_s:3.625 data_s:0.681 smp/s:15 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.001 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.118 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
89
+ Training: 94%|█████████▍| 9400/10000 [10:50:58<59:54, 5.99s/step] INFO 2026-07-24 03:26:29 ot_train.py:661 step:59K smpl:4M ep:86K epch:17.41 loss:0.015 grdn:0.206 lr:3.7e-08 updt_s:3.835 data_s:1.134 smp/s:13 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.000 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.130 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
90
+ Training: 96%|█████████▌| 9600/10000 [11:05:41<28:38, 4.30s/step]INFO 2026-07-24 03:41:13 ot_train.py:661 step:60K smpl:4M ep:86K epch:17.47 loss:0.015 grdn:0.200 lr:1.9e-08 updt_s:3.621 data_s:0.793 smp/s:15 mem_gb:29.85 flow_matching_loss:0.015 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.015 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.000 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.128 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
91
+ Training: 98%|█████████▊| 9800/10000 [11:19:44<14:04, 4.22s/step]INFO 2026-07-24 03:55:15 ot_train.py:661 step:60K smpl:4M ep:86K epch:17.53 loss:0.014 grdn:0.196 lr:7.1e-09 updt_s:3.584 data_s:0.626 smp/s:15 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.000 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.124 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
92
+ Training: 100%|██████████| 10000/10000 [11:33:57<00:00, 3.37s/step]INFO 2026-07-24 04:09:29 ot_train.py:661 step:60K smpl:4M ep:87K epch:17.59 loss:0.014 grdn:0.208 lr:1.0e-09 updt_s:3.553 data_s:0.712 smp/s:15 mem_gb:29.85 flow_matching_loss:0.014 rkd_loss:0.003 rkd_distance_loss:0.001 rkd_angle_loss:0.001 weighted_flow_matching_loss:0.014 weighted_rkd_loss:0.000 rkd_fm_loss_weight:1.000 rkd_loss_weight:0.000 rkd_runtime_enabled:1.000 rkd_student_pooling_id:1.000 rkd_teacher_latent_norm:1.000 rkd_student_feature_norm:15.128 training_phase_index:0.000 training_phase_start_step:0.000 training_phase_until_step:60000.000 training_phase_trainable_params:1080755329.000
93
+ INFO 2026-07-24 04:09:29 ot_train.py:707 Checkpoint policy after step 60000
94
+ INFO 2026-07-24 04:10:06 in_yaml.py:1414 Saved verified rolling checkpoint step=60000 slot=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01
95
+ Training: 100%|██████████| 10000/10000 [11:34:35<00:00, 4.17s/step]
96
+ INFO 2026-07-24 04:10:06 ot_train.py:793 End of training
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/requirements.txt ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ lerobot==0.6.1
2
+ tzdata==2026.3
3
+ nvidia-cufile-cu12==1.13.1.3
4
+ wandb==0.27.2
5
+ six==1.17.0
6
+ peft==0.19.1
7
+ nvidia-curand-cu12==10.3.9.90
8
+ nvidia-cuda-runtime-cu12==12.8.90
9
+ datasets==4.8.5
10
+ attrs==26.1.0
11
+ cloudpickle==3.1.2
12
+ packaging==25.0
13
+ nvidia-cublas-cu12==12.8.4.1
14
+ cmake==4.1.3
15
+ psutil==7.2.2
16
+ pydantic==2.13.4
17
+ GitPython==3.1.51
18
+ pillow==12.2.0
19
+ pytz==2026.2
20
+ cuda-pathfinder==1.2.2
21
+ nvidia-cusparse-cu12==12.5.8.93
22
+ typing_extensions==4.15.0
23
+ urllib3==2.7.0
24
+ pydantic_core==2.46.4
25
+ draccus==0.10.0
26
+ mypy_extensions==1.1.0
27
+ dill==0.4.1
28
+ typer==0.26.8
29
+ decord==0.6.0
30
+ tqdm==4.68.4
31
+ triton==3.6.0
32
+ idna==3.18
33
+ requests==2.34.2
34
+ nvidia-cudnn-cu12==9.19.0.56
35
+ mdurl==0.1.2
36
+ shellingham==1.5.4
37
+ zipp==4.1.0
38
+ timm==1.0.28
39
+ torch==2.11.0+cu128
40
+ tokenizers==0.22.2
41
+ nvidia-cusparselt-cu12==0.7.1
42
+ nvidia-cufft-cu12==11.3.3.83
43
+ nvidia-cusolver-cu12==11.7.3.90
44
+ opencv-python-headless==4.13.0.92
45
+ httpx==0.28.1
46
+ Jinja2==3.1.6
47
+ einops==0.8.2
48
+ mergedeep==1.3.4
49
+ platformdirs==4.10.0
50
+ nvidia-nvjitlink-cu12==12.8.93
51
+ lerobot==0.6.0
52
+ PyYAML==6.0.3
53
+ accelerate==1.14.0
54
+ anyio==4.14.2
55
+ gymnasium==1.3.0
56
+ pip==26.1.2
57
+ huggingface_hub==1.23.0
58
+ diffusers==0.35.2
59
+ annotated-doc==0.0.4
60
+ pyarrow==25.0.0
61
+ python-dateutil==2.9.0.post0
62
+ termcolor==3.3.0
63
+ nvidia-nccl-cu12==2.28.9
64
+ sentry-sdk==2.65.0
65
+ torchcodec==0.11.1
66
+ pyyaml-include==1.4.1
67
+ multiprocess==0.70.19
68
+ wrapt==2.2.2
69
+ typing-inspection==0.4.2
70
+ nvidia-cuda-nvrtc-cu12==12.8.93
71
+ Pygments==2.20.0
72
+ nvidia-cuda-cupti-cu12==12.8.90
73
+ certifi==2026.6.17
74
+ setuptools==78.1.0
75
+ charset-normalizer==3.4.9
76
+ propcache==0.5.2
77
+ multidict==6.7.1
78
+ regex==2026.7.10
79
+ frozenlist==1.8.0
80
+ smmap==5.0.3
81
+ hf-xet==1.5.1
82
+ torchvision==0.26.0+cu128
83
+ gitdb==4.0.12
84
+ transformers==5.5.4
85
+ rich==15.0.0
86
+ cuda-bindings==12.9.4
87
+ protobuf==7.35.1
88
+ aiosignal==1.4.0
89
+ jsonlines==4.0.0
90
+ yarl==1.24.2
91
+ filelock==3.29.0
92
+ av==15.1.0
93
+ networkx==3.6.1
94
+ sympy==1.14.0
95
+ wheel==0.47.0
96
+ nvidia-nvtx-cu12==12.8.90
97
+ xxhash==3.8.1
98
+ toml==0.10.2
99
+ MarkupSafe==3.0.3
100
+ pandas==2.3.3
101
+ safetensors==0.8.0
102
+ click==8.4.2
103
+ fsspec==2026.2.0
104
+ dm-tree==0.1.10
105
+ absl-py==2.5.0
106
+ Farama-Notifications==0.0.6
107
+ mpmath==1.3.0
108
+ importlib_metadata==9.0.0
109
+ markdown-it-py==4.2.0
110
+ h11==0.16.0
111
+ typing-inspect==0.9.0
112
+ httpcore==1.0.9
113
+ annotated-types==0.7.0
114
+ numpy==2.2.6
115
+ nvidia-nvshmem-cu12==3.4.5
116
+ aiohappyeyeballs==2.7.1
117
+ cuda-toolkit==12.8.1
118
+ aiohttp==3.14.1
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/wandb-metadata.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-6.8.0-55-generic-x86_64-with-glibc2.39",
3
+ "python": "CPython 3.12.13",
4
+ "startedAt": "2026-07-23T16:35:01.611693Z",
5
+ "args": [
6
+ "--config_path=/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model/train_config.json",
7
+ "--resume=true"
8
+ ],
9
+ "program": "/home/ext_minje/clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py",
10
+ "codePath": "clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py",
11
+ "codePathLocal": "clvla/benchmarks/INSIGHT/my_scripts/lerobot_train_yaml.py",
12
+ "root": "/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K",
13
+ "host": "ext_minje-box",
14
+ "executable": "/home/ext_minje/miniconda3/envs/lerobot060_groot/bin/python",
15
+ "cpu_count": 112,
16
+ "cpu_count_logical": 224,
17
+ "gpu": "NVIDIA H200",
18
+ "gpu_count": 2,
19
+ "disk": {
20
+ "/": {
21
+ "total": "1888425144320",
22
+ "used": "1327086579712"
23
+ }
24
+ },
25
+ "memory": {
26
+ "total": "2164176568320"
27
+ },
28
+ "gpu_nvidia": [
29
+ {
30
+ "name": "NVIDIA H200",
31
+ "memoryTotal": "150754820096",
32
+ "cudaCores": 16896,
33
+ "architecture": "Hopper",
34
+ "uuid": "GPU-86be6e5b-fb48-fa1c-9bd7-83d336cb2991"
35
+ },
36
+ {
37
+ "name": "NVIDIA H200",
38
+ "memoryTotal": "150754820096",
39
+ "cudaCores": 16896,
40
+ "architecture": "Hopper",
41
+ "uuid": "GPU-24eb1ba2-7313-ff93-1435-748768ebdbdc"
42
+ }
43
+ ],
44
+ "cudaVersion": "12.9",
45
+ "writerId": "3vx7sedbdsz3uay6fuuuh2e3qlqtbnr9"
46
+ }
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/files/wandb-summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"train/training_phase_trainable_params":1.080755329e+09,"train/update_s":3.5525975197848676,"train/rkd_teacher_latent_norm":1,"train/rkd_student_pooling_id":1,"train/weighted_rkd_loss":2.6304401752318693e-08,"train/rkd_runtime_enabled":1,"train/episodes":86694.41811262691,"train/training_phase_index":0,"train/samples_per_s":15.006542098698269,"train/flow_matching_loss":0.014334074733778835,"train/epochs":17.58507466787564,"train/rkd_loss_weight":9.207125008732775e-06,"_runtime":83988,"train/lr":1.0049909226259037e-09,"train/loss":0.014334101057611405,"train/rkd_student_feature_norm":15.128039054870605,"_timestamp":1.7848661692599273e+09,"train/rkd_fm_loss_weight":1,"train/rkd_angle_loss":0.0010985716705908999,"_step":60000,"train/dataloading_s":0.712209094495156,"train/gpu_mem_gb":29.8462087392807,"train/steps":60000,"train/samples":3840000,"train/weighted_flow_matching_loss":0.014334074733778835,"_wandb":{"runtime":83988},"train/training_phase_start_step":0,"train/rkd_loss":0.0028473387891426684,"train/rkd_distance_loss":0.0006501954395207577,"train/training_phase_until_step":60000,"train/grad_norm":0.20803156722337007}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-core.log ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-07-23T16:35:01.682970498Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpa9piztjm/port-1456.txt","pid":1456,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
2
+ {"time":"2026-07-23T16:35:01.682978333Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmplz0alzhu/port-1457.txt","pid":1457,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
3
+ {"time":"2026-07-23T16:35:01.682985044Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpwunx93cw/port-1458.txt","pid":1458,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
4
+ {"time":"2026-07-23T16:35:01.6829908Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpf4abll73/port-1459.txt","pid":1459,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
5
+ {"time":"2026-07-23T16:35:01.68360802Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1456}
6
+ {"time":"2026-07-23T16:35:01.683616029Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1457}
7
+ {"time":"2026-07-23T16:35:01.683634398Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1458}
8
+ {"time":"2026-07-23T16:35:01.683603455Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1456-2506-1311874521/socket","Net":"unix"}}
9
+ {"time":"2026-07-23T16:35:01.68361302Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1457-2507-772500566/socket","Net":"unix"}}
10
+ {"time":"2026-07-23T16:35:01.683630442Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1458-2509-4014368169/socket","Net":"unix"}}
11
+ {"time":"2026-07-23T16:35:01.683627313Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1459-2508-1604368573/socket","Net":"unix"}}
12
+ {"time":"2026-07-23T16:35:01.683696179Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1459}
13
+ {"time":"2026-07-23T16:35:01.832927644Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
14
+ {"time":"2026-07-23T16:35:01.833171958Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
15
+ {"time":"2026-07-23T16:35:01.833208155Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
16
+ {"time":"2026-07-23T16:35:01.833275727Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
17
+ {"time":"2026-07-23T16:35:01.837817931Z","level":"INFO","msg":"handleInformInit: received","streamId":"rs1egwv8","id":"1(@)"}
18
+ {"time":"2026-07-23T16:35:01.837959217Z","level":"INFO","msg":"handleInformInit: received","streamId":"2vl3enlf","id":"1(@)"}
19
+ {"time":"2026-07-23T16:35:01.838627327Z","level":"INFO","msg":"handleInformInit: received","streamId":"jeqgwsec","id":"1(@)"}
20
+ {"time":"2026-07-23T16:35:01.838419673Z","level":"INFO","msg":"handleInformInit: received","streamId":"x6da8gup","id":"1(@)"}
21
+ {"time":"2026-07-23T16:35:01.918297888Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpuyiti7pv/port-1460.txt","pid":1460,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
22
+ {"time":"2026-07-23T16:35:01.91863949Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1460}
23
+ {"time":"2026-07-23T16:35:01.918628695Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1460-2829-1778593576/socket","Net":"unix"}}
24
+ {"time":"2026-07-23T16:35:02.074586625Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
25
+ {"time":"2026-07-23T16:35:02.079634061Z","level":"INFO","msg":"handleInformInit: received","streamId":"brma7856","id":"1(@)"}
26
+ {"time":"2026-07-23T16:35:02.303790584Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"2vl3enlf","id":"1(@)"}
27
+ {"time":"2026-07-23T16:35:02.30665649Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"x6da8gup","id":"1(@)"}
28
+ {"time":"2026-07-23T16:35:02.307741694Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"jeqgwsec","id":"1(@)"}
29
+ {"time":"2026-07-23T16:35:02.32303733Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"rs1egwv8","id":"1(@)"}
30
+ {"time":"2026-07-23T16:35:02.493177984Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"brma7856","id":"1(@)"}
31
+ {"time":"2026-07-23T16:35:08.314840933Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"pomu74zdcjwn"}
32
+ {"time":"2026-07-23T16:35:08.415606838Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"u8kbql7lf4bh"}
33
+ {"time":"2026-07-23T16:35:08.420996327Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"bhwn826heh2c"}
34
+ {"time":"2026-07-23T16:35:08.623677864Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"0dnh3nypnhfm"}
35
+ {"time":"2026-07-23T16:35:09.044152431Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"4z6b40zq3yfu"}
36
+ {"time":"2026-07-23T18:11:27.126647161Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
37
+ {"time":"2026-07-23T18:11:27.126779892Z","level":"INFO","msg":"server is shutting down"}
38
+ {"time":"2026-07-23T18:11:27.12683822Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-1457-2507-772500566/socket","Net":"unix"}}
39
+ {"time":"2026-07-23T18:11:27.126852003Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
40
+ {"time":"2026-07-23T18:11:27.126886112Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
41
+ {"time":"2026-07-23T18:11:27.126932287Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
42
+ {"time":"2026-07-23T18:11:27.378619144Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
43
+ {"time":"2026-07-23T18:11:27.37874002Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
44
+ {"time":"2026-07-23T18:11:27.378802617Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
45
+ {"time":"2026-07-23T18:11:27.378812716Z","level":"INFO","msg":"server is shutting down"}
46
+ {"time":"2026-07-23T18:11:27.378878838Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-1456-2506-1311874521/socket","Net":"unix"}}
47
+ {"time":"2026-07-23T18:11:27.378895674Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
48
+ {"time":"2026-07-23T18:11:35.494138091Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
49
+ {"time":"2026-07-23T18:11:35.494189389Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
50
+ {"time":"2026-07-23T18:11:35.494201536Z","level":"INFO","msg":"server is closed"}
51
+ {"time":"2026-07-23T18:12:28.994694228Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
52
+ {"time":"2026-07-23T18:12:28.994728432Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
53
+ {"time":"2026-07-23T18:12:28.994736061Z","level":"INFO","msg":"server is closed"}
54
+ {"time":"2026-07-24T04:08:45.417129571Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
55
+ {"time":"2026-07-24T04:08:45.417350121Z","level":"INFO","msg":"server is shutting down"}
56
+ {"time":"2026-07-24T04:08:45.417341706Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
57
+ {"time":"2026-07-24T04:08:45.417373499Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
58
+ {"time":"2026-07-24T04:08:45.41745467Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
59
+ {"time":"2026-07-24T04:08:45.417448026Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-1458-2509-4014368169/socket","Net":"unix"}}
60
+ {"time":"2026-07-24T04:08:47.515259401Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
61
+ {"time":"2026-07-24T04:08:47.515290273Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
62
+ {"time":"2026-07-24T04:08:47.515301151Z","level":"INFO","msg":"server is closed"}
63
+ {"time":"2026-07-24T04:10:06.98288621Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
64
+ {"time":"2026-07-24T04:10:06.983062676Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
65
+ {"time":"2026-07-24T04:10:06.983136342Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
66
+ {"time":"2026-07-24T04:10:06.983144044Z","level":"INFO","msg":"server is shutting down"}
67
+ {"time":"2026-07-24T04:10:06.98322588Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-1459-2508-1604368573/socket","Net":"unix"}}
68
+ {"time":"2026-07-24T04:10:06.983271294Z","level":"INFO","msg":"processOutgoingData: finished","id":"1(@)"}
69
+ {"time":"2026-07-24T04:10:23.060371611Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
70
+ {"time":"2026-07-24T04:10:23.060399527Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
71
+ {"time":"2026-07-24T04:10:23.060415126Z","level":"INFO","msg":"server is closed"}
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-internal.log ADDED
The diff for this file is too large to render. See raw diff
 
Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug.log ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Current SDK version is 0.27.2
2
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Configure stats pid to 1459
3
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug.log
5
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/wandb/run-20260723_163501-rs1egwv8/logs/debug-internal.log
6
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():766] calling init triggers
7
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():771] wandb.init called with sweep_config: {}
8
+ config: {'dataset': {'repo_id': 'Whalswp/INSIGHTfixposV4_filtered_multispace_v2', 'root': '/home/ext_minje/INSIGHTfixposV4_filtered_multispace_v2', 'episodes': None, 'image_transforms': {'enable': False, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 1.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 1.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 1.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'torchcodec', 'return_uint8': False, 'depth_output_unit': 'mm', 'streaming': False, 'eval_split': 0.0}, 'env': None, 'policy': {'type': 'groot_rkd', 'n_obs_steps': 1, 'input_features': {'observation.state': {'type': <FeatureType.STATE: 'STATE'>, 'shape': [16]}, 'observation.images.wrist': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.right_shoulder': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}, 'observation.images.guide': {'type': <FeatureType.VISUAL: 'VISUAL'>, 'shape': [3, 224, 224]}}, 'output_features': {'action': {'type': <FeatureType.ACTION: 'ACTION'>, 'shape': [10]}}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': None, 'private': None, 'tags': None, 'license': None, 'pretrained_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01/pretrained_model', 'pretrained_revision': None, 'chunk_size': 16, 'n_action_steps': 16, 'max_state_dim': 132, 'max_action_dim': 132, 'normalization_mapping': {'VISUAL': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'STATE': <NormalizationMode.IDENTITY: 'IDENTITY'>, 'ACTION': <NormalizationMode.IDENTITY: 'IDENTITY'>}, 'base_model_path': 'nvidia/GR00T-N1.7-3B', 'action_decode_transform': None, 'embodiment_tag': 'new_embodiment', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'tune_vlln': True, 'tune_top_llm_layers': 0, 'num_inference_timesteps': None, 'rtc_ramp_rate': None, 'use_flash_attention': False, 'use_relative_actions': False, 'relative_exclude_joints': [], 'optimizer_lr': 0.0001, 'optimizer_betas': [0.9, 0.999], 'optimizer_eps': 1e-08, 'optimizer_weight_decay': 1e-05, 'warmup_steps': 3000, 'warmup_ratio': 0.05, 'use_bf16': True, 'model_params_fp32': True, 'image_size': [256, 256], 'tokenizer_assets_repo': None, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_full_model': False, 'video_backend': 'decord', 'balance_dataset_weights': True, 'balance_trajectory_weights': True, 'dataset_paths': None, 'output_dir': './tmp/gr00t', 'save_steps': 1000, 'max_steps': 10000, 'batch_size': 32, 'dataloader_num_workers': 8, 'report_to': 'wandb', 'resume': False, 'action_expert_num_layers': 16, 'rkd_enabled': True, 'rkd_vae_checkpoint': '/home/ext_minje/groot_insight/TimewarpVAE_ckpt/TimewarpVAE_combo_top8_30k/C8_01_C3_008_beta_5em05_latent_dim_16_granularity_4_decoder_hidden_dim_512/checkpoint-step-30000-epoch-39.11', 'rkd_student_pooling': 'attention', 'rkd_student_attention_hidden_dim': 512, 'rkd_student_attention_output_dim': 512, 'rkd_fm_loss_weight': 1.0, 'rkd_loss_weight': 1.0, 'rkd_distance_weight': 1.0, 'rkd_angle_weight': 2.0}, 'reward_model': None, 'output_dir': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K', 'job_name': 'INSIGHT_6D_DiT16_RKD_TimewarpVAE_RSCLstyle_C60', 'resume': True, 'seed': 42, 'cudnn_deterministic': False, 'num_workers': 4, 'batch_size': 64, 'prefetch_factor': 4, 'persistent_workers': True, 'steps': 60000, 'env_eval_freq': 20000, 'log_freq': 200, 'eval_steps': 0, 'max_eval_samples': 0, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 20000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 0.0001, 'weight_decay': 1e-05, 'grad_clip_norm': 1.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': {'type': 'diffuser', 'num_warmup_steps': 3000, 'name': 'cosine'}, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': True, 'recording': False, 'recording_repo_id': None, 'recording_private': False}, 'wandb': {'enable': True, 'disable_artifact': True, 'project': 'lerobot', 'entity': None, 'notes': None, 'run_id': 'rs1egwv8', 'mode': None, 'add_tags': True}, 'peft': None, 'job': {'target': None, 'image': 'huggingface/lerobot-gpu:latest', 'timeout': '2d', 'detach': False, 'tags': []}, 'save_checkpoint_to_hub': False, 'sample_weighting': None, 'rename_map': {}, 'checkpoint_path': '/home/ext_minje/groot_insight/train_ckpt/Abs_6D/DiT_Layer/16/RSCLstyle_cosine_60K/checkpoints/rolling_01', '_wandb': {}}
9
+ 2026-07-23 16:35:01,612 INFO MainThread:1459 [wandb_init.py:init():814] starting backend
10
+ 2026-07-23 16:35:01,833 INFO MainThread:1459 [wandb_init.py:init():829] sending inform_init request
11
+ 2026-07-23 16:35:02,323 INFO MainThread:1459 [wandb_init.py:init():834] backend started and connected
12
+ 2026-07-23 16:35:02,325 INFO MainThread:1459 [wandb_init.py:init():904] updated telemetry
13
+ 2026-07-23 16:35:02,326 INFO MainThread:1459 [wandb_init.py:init():927] communicating run to backend with 90.0 second timeout
14
+ 2026-07-23 16:35:03,342 INFO MainThread:1459 [wandb_init.py:init():967] run resumed
15
+ 2026-07-23 16:35:03,344 INFO MainThread:1459 [wandb_init.py:init():972] starting run threads in backend
16
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_console_start():2523] atexit reg
17
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2373] redirect: wrap_raw
18
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2442] Wrapping output streams.
19
+ 2026-07-23 16:35:03,396 INFO MainThread:1459 [wandb_run.py:_redirect():2465] Redirects installed.
20
+ 2026-07-23 16:35:03,398 INFO MainThread:1459 [wandb_init.py:init():1010] run started, returning control to user process
21
+ 2026-07-24 04:10:06,982 INFO wandb-AsyncioManager-main:1459 [service_client.py:_forward_responses():122] Reached EOF.
22
+ 2026-07-24 04:10:06,983 INFO wandb-AsyncioManager-main:1459 [mailbox.py:close():154] Closing mailbox, abandoning 2 handles.
23
+ 2026-07-24 04:10:06,983 ERROR wandb-AsyncioManager-main:1459 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
24
+ Traceback (most recent call last):
25
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
26
+ await fn()
27
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 90, in loop
28
+ await asyncio_compat.race(
29
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 278, in race
30
+ async with open_task_group(race=True) as tg:
31
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^
32
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/contextlib.py", line 217, in __aexit__
33
+ await anext(self.gen)
34
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 240, in open_task_group
35
+ await task_group._wait_all(race=race, timeout=exit_timeout)
36
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_compat.py", line 180, in _wait_all
37
+ raise exc
38
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_messages.py", line 113, in _print_all
39
+ result = await handle.wait_async(timeout=None)
40
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
41
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox_handle.py", line 127, in wait_async
42
+ response = await self._handle.wait_async(timeout=timeout)
43
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
44
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/response_handle.py", line 123, in wait_async
45
+ raise HandleAbandonedError()
46
+ wandb.sdk.mailbox.mailbox_handle.HandleAbandonedError
47
+ 2026-07-24 04:10:14,759 ERROR wandb-AsyncioManager-main:1459 [asyncio_manager.py:fn_wrap_exceptions():184] Uncaught exception in run_soon callback.
48
+ Traceback (most recent call last):
49
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 182, in fn_wrap_exceptions
50
+ await fn()
51
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 87, in loop
52
+ should_stop = await self._check_should_stop()
53
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
54
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/run_stopping.py", line 115, in _check_should_stop
55
+ handle = await self._interface.deliver_async(request)
56
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
57
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/interface/interface_sock.py", line 58, in deliver_async
58
+ handle = await self._client.deliver(request)
59
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
60
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 57, in deliver
61
+ handle = self._mailbox.require_response(request)
62
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
63
+ File "/home/ext_minje/miniconda3/envs/lerobot060_groot/lib/python3.12/site-packages/wandb/sdk/mailbox/mailbox.py", line 89, in require_response
64
+ raise MailboxClosedError()
65
+ wandb.sdk.mailbox.mailbox.MailboxClosedError