Whalswp commited on
Commit
f5965cd
·
verified ·
1 Parent(s): 53b6768

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +2 -0
  2. checkpoints/050000/pretrained_model/config.json +103 -0
  3. checkpoints/050000/pretrained_model/model.safetensors +3 -0
  4. checkpoints/050000/pretrained_model/policy_postprocessor.json +20 -0
  5. checkpoints/050000/pretrained_model/policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors +3 -0
  6. checkpoints/050000/pretrained_model/policy_preprocessor.json +52 -0
  7. checkpoints/050000/pretrained_model/policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors +3 -0
  8. checkpoints/050000/pretrained_model/train_config.json +258 -0
  9. checkpoints/050000/training_state/optimizer_param_groups.json +335 -0
  10. checkpoints/050000/training_state/optimizer_state.safetensors +3 -0
  11. checkpoints/050000/training_state/rng_state.safetensors +3 -0
  12. checkpoints/050000/training_state/scheduler_state.json +14 -0
  13. checkpoints/050000/training_state/training_step.json +3 -0
  14. checkpoints/100000/pretrained_model/config.json +103 -0
  15. checkpoints/100000/pretrained_model/model.safetensors +3 -0
  16. checkpoints/100000/pretrained_model/policy_postprocessor.json +20 -0
  17. checkpoints/100000/pretrained_model/policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors +3 -0
  18. checkpoints/100000/pretrained_model/policy_preprocessor.json +52 -0
  19. checkpoints/100000/pretrained_model/policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors +3 -0
  20. checkpoints/100000/pretrained_model/train_config.json +258 -0
  21. checkpoints/100000/training_state/optimizer_param_groups.json +335 -0
  22. checkpoints/100000/training_state/optimizer_state.safetensors +3 -0
  23. checkpoints/100000/training_state/rng_state.safetensors +3 -0
  24. checkpoints/100000/training_state/scheduler_state.json +14 -0
  25. checkpoints/100000/training_state/training_step.json +3 -0
  26. dataset_sample_report.json +43 -0
  27. first_batch_report.json +207 -0
  28. pipeline_commands.json +65 -0
  29. pipeline_manifest.json +91 -0
  30. run_manifest.json +114 -0
  31. train_summary_report.json +17 -0
  32. trainable_param_report.json +28 -0
  33. wandb/debug-internal.log +41 -0
  34. wandb/debug.log +24 -0
  35. wandb/run-20260518_152847-smy9mkg4/files/config.yaml +178 -0
  36. wandb/run-20260518_152847-smy9mkg4/files/output.log +103 -0
  37. wandb/run-20260518_152847-smy9mkg4/files/requirements.txt +209 -0
  38. wandb/run-20260518_152847-smy9mkg4/files/wandb-metadata.json +87 -0
  39. wandb/run-20260518_152847-smy9mkg4/files/wandb-summary.json +1 -0
  40. wandb/run-20260518_152847-smy9mkg4/logs/debug-core.log +14 -0
  41. wandb/run-20260518_152847-smy9mkg4/logs/debug-internal.log +11 -0
  42. wandb/run-20260518_152847-smy9mkg4/logs/debug.log +34 -0
  43. wandb/run-20260518_152847-smy9mkg4/run-smy9mkg4.wandb +0 -0
  44. wandb/run-20260518_153314-yn3lgfkj/files/config.yaml +178 -0
  45. wandb/run-20260518_153314-yn3lgfkj/files/output.log +3 -0
  46. wandb/run-20260518_153314-yn3lgfkj/files/requirements.txt +209 -0
  47. wandb/run-20260518_153314-yn3lgfkj/files/wandb-metadata.json +87 -0
  48. wandb/run-20260518_153314-yn3lgfkj/files/wandb-summary.json +1 -0
  49. wandb/run-20260518_153314-yn3lgfkj/logs/debug-core.log +12 -0
  50. wandb/run-20260518_153314-yn3lgfkj/logs/debug-internal.log +11 -0
.gitattributes CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ wandb/run-20260518_153916-gnre0khg/run-gnre0khg.wandb filter=lfs diff=lfs merge=lfs -text
37
+ wandb/run-20260519_221333-qz452p60/run-qz452p60.wandb filter=lfs diff=lfs merge=lfs -text
checkpoints/050000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "groot",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 16
9
+ ]
10
+ },
11
+ "observation.images.guide": {
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.wrist": {
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
+ 8
41
+ ]
42
+ },
43
+ "action.skill_id": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 1
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": true,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "chunk_size": 16,
60
+ "n_action_steps": 16,
61
+ "max_state_dim": 64,
62
+ "max_action_dim": 32,
63
+ "normalization_mapping": {
64
+ "VISUAL": "IDENTITY",
65
+ "STATE": "MEAN_STD",
66
+ "ACTION": "MEAN_STD"
67
+ },
68
+ "image_size": [
69
+ 224,
70
+ 224
71
+ ],
72
+ "base_model_path": "nvidia/GR00T-N1.5-3B",
73
+ "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
74
+ "embodiment_tag": "new_embodiment",
75
+ "tune_llm": false,
76
+ "tune_visual": false,
77
+ "tune_projector": true,
78
+ "tune_diffusion_model": true,
79
+ "lora_rank": 0,
80
+ "lora_alpha": 16,
81
+ "lora_dropout": 0.1,
82
+ "lora_full_model": false,
83
+ "optimizer_lr": 0.0001,
84
+ "optimizer_betas": [
85
+ 0.95,
86
+ 0.999
87
+ ],
88
+ "optimizer_eps": 1e-08,
89
+ "optimizer_weight_decay": 1e-05,
90
+ "warmup_ratio": 0.05,
91
+ "use_bf16": true,
92
+ "video_backend": "pyav",
93
+ "balance_dataset_weights": true,
94
+ "balance_trajectory_weights": true,
95
+ "dataset_paths": null,
96
+ "output_dir": "./tmp/gr00t",
97
+ "save_steps": 1000,
98
+ "max_steps": 10000,
99
+ "batch_size": 32,
100
+ "dataloader_num_workers": 8,
101
+ "report_to": "wandb",
102
+ "resume": false
103
+ }
checkpoints/050000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:224dc1aa459dcd4f0077260e508b34fef13f3555b3c9912ea9fbee3a787ca76c
3
+ size 5448472032
checkpoints/050000/pretrained_model/policy_postprocessor.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "groot_action_unpack_unnormalize_v1",
6
+ "config": {
7
+ "env_action_dim": 8,
8
+ "normalize_min_max": true
9
+ },
10
+ "state_file": "policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors"
11
+ },
12
+ {
13
+ "registry_name": "device_processor",
14
+ "config": {
15
+ "device": "cpu",
16
+ "float_dtype": null
17
+ }
18
+ }
19
+ ]
20
+ }
checkpoints/050000/pretrained_model/policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cacc0ce3063ffe52d5e066d64ce32242b357c4ba7694ecc40f9937b7ca6a95a4
3
+ size 15424
checkpoints/050000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {}
8
+ }
9
+ },
10
+ {
11
+ "registry_name": "to_batch_processor",
12
+ "config": {}
13
+ },
14
+ {
15
+ "registry_name": "groot_pack_inputs_v3",
16
+ "config": {
17
+ "state_horizon": 1,
18
+ "action_horizon": 16,
19
+ "max_state_dim": 64,
20
+ "max_action_dim": 32,
21
+ "language_key": "task",
22
+ "formalize_language": false,
23
+ "embodiment_tag": "new_embodiment",
24
+ "embodiment_mapping": {
25
+ "new_embodiment": 31,
26
+ "oxe_droid": 17,
27
+ "agibot_genie1": 26,
28
+ "gr1": 24,
29
+ "so100": 2,
30
+ "unitree_g1": 3
31
+ },
32
+ "normalize_min_max": true
33
+ },
34
+ "state_file": "policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors"
35
+ },
36
+ {
37
+ "registry_name": "groot_eagle_encode_v3",
38
+ "config": {}
39
+ },
40
+ {
41
+ "registry_name": "groot_eagle_collate_v3",
42
+ "config": {}
43
+ },
44
+ {
45
+ "registry_name": "device_processor",
46
+ "config": {
47
+ "device": "cuda",
48
+ "float_dtype": null
49
+ }
50
+ }
51
+ ]
52
+ }
checkpoints/050000/pretrained_model/policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cacc0ce3063ffe52d5e066d64ce32242b357c4ba7694ecc40f9937b7ca6a95a4
3
+ size 15424
checkpoints/050000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "/home/seonho/insight/insight_dataset",
4
+ "root": "/home/seonho/insight/insight_dataset",
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": "pyav",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "groot",
85
+ "n_obs_steps": 1,
86
+ "input_features": {
87
+ "observation.state": {
88
+ "type": "STATE",
89
+ "shape": [
90
+ 16
91
+ ]
92
+ },
93
+ "observation.images.guide": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 3,
97
+ 224,
98
+ 224
99
+ ]
100
+ },
101
+ "observation.images.right_shoulder": {
102
+ "type": "VISUAL",
103
+ "shape": [
104
+ 3,
105
+ 224,
106
+ 224
107
+ ]
108
+ },
109
+ "observation.images.wrist": {
110
+ "type": "VISUAL",
111
+ "shape": [
112
+ 3,
113
+ 224,
114
+ 224
115
+ ]
116
+ }
117
+ },
118
+ "output_features": {
119
+ "action": {
120
+ "type": "ACTION",
121
+ "shape": [
122
+ 8
123
+ ]
124
+ },
125
+ "action.skill_id": {
126
+ "type": "ACTION",
127
+ "shape": [
128
+ 1
129
+ ]
130
+ }
131
+ },
132
+ "device": "cuda",
133
+ "use_amp": false,
134
+ "use_peft": false,
135
+ "push_to_hub": true,
136
+ "repo_id": null,
137
+ "private": null,
138
+ "tags": null,
139
+ "license": null,
140
+ "pretrained_path": null,
141
+ "chunk_size": 16,
142
+ "n_action_steps": 16,
143
+ "max_state_dim": 64,
144
+ "max_action_dim": 32,
145
+ "normalization_mapping": {
146
+ "VISUAL": "IDENTITY",
147
+ "STATE": "MEAN_STD",
148
+ "ACTION": "MEAN_STD"
149
+ },
150
+ "image_size": [
151
+ 224,
152
+ 224
153
+ ],
154
+ "base_model_path": "nvidia/GR00T-N1.5-3B",
155
+ "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
156
+ "embodiment_tag": "new_embodiment",
157
+ "tune_llm": false,
158
+ "tune_visual": false,
159
+ "tune_projector": true,
160
+ "tune_diffusion_model": true,
161
+ "lora_rank": 0,
162
+ "lora_alpha": 16,
163
+ "lora_dropout": 0.1,
164
+ "lora_full_model": false,
165
+ "optimizer_lr": 0.0001,
166
+ "optimizer_betas": [
167
+ 0.95,
168
+ 0.999
169
+ ],
170
+ "optimizer_eps": 1e-08,
171
+ "optimizer_weight_decay": 1e-05,
172
+ "warmup_ratio": 0.05,
173
+ "use_bf16": true,
174
+ "video_backend": "pyav",
175
+ "balance_dataset_weights": true,
176
+ "balance_trajectory_weights": true,
177
+ "dataset_paths": null,
178
+ "output_dir": "./tmp/gr00t",
179
+ "save_steps": 1000,
180
+ "max_steps": 10000,
181
+ "batch_size": 32,
182
+ "dataloader_num_workers": 8,
183
+ "report_to": "wandb",
184
+ "resume": false
185
+ },
186
+ "output_dir": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
187
+ "job_name": "groot_insight",
188
+ "resume": false,
189
+ "seed": 42,
190
+ "cudnn_deterministic": false,
191
+ "num_workers": 8,
192
+ "batch_size": 64,
193
+ "steps": 100000,
194
+ "eval_freq": 20000,
195
+ "log_freq": 100,
196
+ "tolerance_s": 0.0001,
197
+ "save_checkpoint": true,
198
+ "save_freq": 50000,
199
+ "use_policy_training_preset": true,
200
+ "optimizer": null,
201
+ "scheduler": null,
202
+ "eval": {
203
+ "n_episodes": 50,
204
+ "batch_size": 50,
205
+ "use_async_envs": false,
206
+ "max_episodes_rendered": 10
207
+ },
208
+ "wandb": {
209
+ "enable": true,
210
+ "disable_artifact": false,
211
+ "project": "groot_insight_baseline_base",
212
+ "entity": "RwHlabs",
213
+ "notes": null,
214
+ "run_id": null,
215
+ "mode": null,
216
+ "add_tags": true,
217
+ "name": "insight_phase1_dit_only_groot",
218
+ "group": "insight_phase1_dit_only",
219
+ "tags": "base,phase2,groot,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
220
+ },
221
+ "peft": null,
222
+ "use_rabc": false,
223
+ "rabc_progress_path": null,
224
+ "rabc_kappa": 0.01,
225
+ "rabc_epsilon": 1e-06,
226
+ "rabc_head_mode": "sparse",
227
+ "rename_map": {},
228
+ "checkpoint_path": null,
229
+ "method": "base",
230
+ "phase": "phase2",
231
+ "pretrained_path": "",
232
+ "gradient_checkpointing": false,
233
+ "smoke_test": false,
234
+ "dry_run": false,
235
+ "grad_clip_norm": 10.0,
236
+ "write_reports": true,
237
+ "fail_on_language_fallback": true,
238
+ "action_dim": 8,
239
+ "state_dim": 16,
240
+ "camera_set": "rgb",
241
+ "video_feature_keys": "",
242
+ "task_prompt_mode": "non_guide",
243
+ "task_descriptions_path": "",
244
+ "excluded_indices_path": "",
245
+ "exclude_index_scope": "none",
246
+ "recipe_yaml": "",
247
+ "pipeline_yaml": "",
248
+ "pipeline_mode": "both",
249
+ "dry_print_commands": false,
250
+ "launcher": "python",
251
+ "phase2_output_dir": "",
252
+ "phase3_output_dir": "",
253
+ "wandb_auto_name": true,
254
+ "run_name": "",
255
+ "wandb_project_prefix": "groot_insight",
256
+ "wandb_group": "",
257
+ "wandb_tags": ""
258
+ }
checkpoints/050000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,335 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 5.4128967273616625e-05,
4
+ "betas": [
5
+ 0.95,
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
+ 159,
179
+ 160,
180
+ 161,
181
+ 162,
182
+ 163,
183
+ 164,
184
+ 165,
185
+ 166,
186
+ 167,
187
+ 168,
188
+ 169,
189
+ 170,
190
+ 171,
191
+ 172,
192
+ 173,
193
+ 174,
194
+ 175,
195
+ 176,
196
+ 177,
197
+ 178,
198
+ 179,
199
+ 180,
200
+ 181,
201
+ 182,
202
+ 183,
203
+ 184,
204
+ 185,
205
+ 186,
206
+ 187,
207
+ 188,
208
+ 189,
209
+ 190,
210
+ 191,
211
+ 192,
212
+ 193,
213
+ 194,
214
+ 195,
215
+ 196,
216
+ 197,
217
+ 198,
218
+ 199,
219
+ 200,
220
+ 201,
221
+ 202,
222
+ 203,
223
+ 204,
224
+ 205,
225
+ 206,
226
+ 207,
227
+ 208,
228
+ 209,
229
+ 210,
230
+ 211,
231
+ 212,
232
+ 213,
233
+ 214,
234
+ 215,
235
+ 216,
236
+ 217,
237
+ 218,
238
+ 219,
239
+ 220,
240
+ 221,
241
+ 222,
242
+ 223,
243
+ 224,
244
+ 225,
245
+ 226,
246
+ 227,
247
+ 228,
248
+ 229,
249
+ 230,
250
+ 231,
251
+ 232,
252
+ 233,
253
+ 234,
254
+ 235,
255
+ 236,
256
+ 237,
257
+ 238,
258
+ 239,
259
+ 240,
260
+ 241,
261
+ 242,
262
+ 243,
263
+ 244,
264
+ 245,
265
+ 246,
266
+ 247,
267
+ 248,
268
+ 249,
269
+ 250,
270
+ 251,
271
+ 252,
272
+ 253,
273
+ 254,
274
+ 255,
275
+ 256,
276
+ 257,
277
+ 258,
278
+ 259,
279
+ 260,
280
+ 261,
281
+ 262,
282
+ 263,
283
+ 264,
284
+ 265,
285
+ 266,
286
+ 267,
287
+ 268,
288
+ 269,
289
+ 270,
290
+ 271,
291
+ 272,
292
+ 273,
293
+ 274,
294
+ 275,
295
+ 276,
296
+ 277,
297
+ 278,
298
+ 279,
299
+ 280,
300
+ 281,
301
+ 282,
302
+ 283,
303
+ 284,
304
+ 285,
305
+ 286,
306
+ 287,
307
+ 288,
308
+ 289,
309
+ 290,
310
+ 291,
311
+ 292,
312
+ 293,
313
+ 294,
314
+ 295,
315
+ 296,
316
+ 297,
317
+ 298,
318
+ 299,
319
+ 300,
320
+ 301,
321
+ 302,
322
+ 303,
323
+ 304,
324
+ 305,
325
+ 306,
326
+ 307,
327
+ 308,
328
+ 309,
329
+ 310,
330
+ 311,
331
+ 312,
332
+ 313
333
+ ]
334
+ }
335
+ ]
checkpoints/050000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b7d4f28f0f9036678bfc35f6175e2eec7348050ca406eeb2e8d41a6b4593a56
3
+ size 4275306416
checkpoints/050000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ddf10679b738bfb52bcb08b700d0fdb5fa15c0aaff47068661cfe7bb427cf9ba
3
+ size 15708
checkpoints/050000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 0.0001
4
+ ],
5
+ "last_epoch": 50000,
6
+ "_step_count": 50001,
7
+ "_get_lr_called_within_step": false,
8
+ "_last_lr": [
9
+ 5.4128967273616625e-05
10
+ ],
11
+ "lr_lambdas": [
12
+ {}
13
+ ]
14
+ }
checkpoints/050000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 50000
3
+ }
checkpoints/100000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "groot",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 16
9
+ ]
10
+ },
11
+ "observation.images.guide": {
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.wrist": {
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
+ 8
41
+ ]
42
+ },
43
+ "action.skill_id": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 1
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": true,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "chunk_size": 16,
60
+ "n_action_steps": 16,
61
+ "max_state_dim": 64,
62
+ "max_action_dim": 32,
63
+ "normalization_mapping": {
64
+ "VISUAL": "IDENTITY",
65
+ "STATE": "MEAN_STD",
66
+ "ACTION": "MEAN_STD"
67
+ },
68
+ "image_size": [
69
+ 224,
70
+ 224
71
+ ],
72
+ "base_model_path": "nvidia/GR00T-N1.5-3B",
73
+ "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
74
+ "embodiment_tag": "new_embodiment",
75
+ "tune_llm": false,
76
+ "tune_visual": false,
77
+ "tune_projector": true,
78
+ "tune_diffusion_model": true,
79
+ "lora_rank": 0,
80
+ "lora_alpha": 16,
81
+ "lora_dropout": 0.1,
82
+ "lora_full_model": false,
83
+ "optimizer_lr": 0.0001,
84
+ "optimizer_betas": [
85
+ 0.95,
86
+ 0.999
87
+ ],
88
+ "optimizer_eps": 1e-08,
89
+ "optimizer_weight_decay": 1e-05,
90
+ "warmup_ratio": 0.05,
91
+ "use_bf16": true,
92
+ "video_backend": "pyav",
93
+ "balance_dataset_weights": true,
94
+ "balance_trajectory_weights": true,
95
+ "dataset_paths": null,
96
+ "output_dir": "./tmp/gr00t",
97
+ "save_steps": 1000,
98
+ "max_steps": 10000,
99
+ "batch_size": 32,
100
+ "dataloader_num_workers": 8,
101
+ "report_to": "wandb",
102
+ "resume": false
103
+ }
checkpoints/100000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c06fb1f18b720c7d5870fd740a4bd777d06f04d826c9bc483254e643c296318
3
+ size 5448472032
checkpoints/100000/pretrained_model/policy_postprocessor.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "groot_action_unpack_unnormalize_v1",
6
+ "config": {
7
+ "env_action_dim": 8,
8
+ "normalize_min_max": true
9
+ },
10
+ "state_file": "policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors"
11
+ },
12
+ {
13
+ "registry_name": "device_processor",
14
+ "config": {
15
+ "device": "cpu",
16
+ "float_dtype": null
17
+ }
18
+ }
19
+ ]
20
+ }
checkpoints/100000/pretrained_model/policy_postprocessor_step_0_groot_action_unpack_unnormalize_v1.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cacc0ce3063ffe52d5e066d64ce32242b357c4ba7694ecc40f9937b7ca6a95a4
3
+ size 15424
checkpoints/100000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {}
8
+ }
9
+ },
10
+ {
11
+ "registry_name": "to_batch_processor",
12
+ "config": {}
13
+ },
14
+ {
15
+ "registry_name": "groot_pack_inputs_v3",
16
+ "config": {
17
+ "state_horizon": 1,
18
+ "action_horizon": 16,
19
+ "max_state_dim": 64,
20
+ "max_action_dim": 32,
21
+ "language_key": "task",
22
+ "formalize_language": false,
23
+ "embodiment_tag": "new_embodiment",
24
+ "embodiment_mapping": {
25
+ "new_embodiment": 31,
26
+ "oxe_droid": 17,
27
+ "agibot_genie1": 26,
28
+ "gr1": 24,
29
+ "so100": 2,
30
+ "unitree_g1": 3
31
+ },
32
+ "normalize_min_max": true
33
+ },
34
+ "state_file": "policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors"
35
+ },
36
+ {
37
+ "registry_name": "groot_eagle_encode_v3",
38
+ "config": {}
39
+ },
40
+ {
41
+ "registry_name": "groot_eagle_collate_v3",
42
+ "config": {}
43
+ },
44
+ {
45
+ "registry_name": "device_processor",
46
+ "config": {
47
+ "device": "cuda",
48
+ "float_dtype": null
49
+ }
50
+ }
51
+ ]
52
+ }
checkpoints/100000/pretrained_model/policy_preprocessor_step_2_groot_pack_inputs_v3.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cacc0ce3063ffe52d5e066d64ce32242b357c4ba7694ecc40f9937b7ca6a95a4
3
+ size 15424
checkpoints/100000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "/home/seonho/insight/insight_dataset",
4
+ "root": "/home/seonho/insight/insight_dataset",
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": "pyav",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "groot",
85
+ "n_obs_steps": 1,
86
+ "input_features": {
87
+ "observation.state": {
88
+ "type": "STATE",
89
+ "shape": [
90
+ 16
91
+ ]
92
+ },
93
+ "observation.images.guide": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 3,
97
+ 224,
98
+ 224
99
+ ]
100
+ },
101
+ "observation.images.right_shoulder": {
102
+ "type": "VISUAL",
103
+ "shape": [
104
+ 3,
105
+ 224,
106
+ 224
107
+ ]
108
+ },
109
+ "observation.images.wrist": {
110
+ "type": "VISUAL",
111
+ "shape": [
112
+ 3,
113
+ 224,
114
+ 224
115
+ ]
116
+ }
117
+ },
118
+ "output_features": {
119
+ "action": {
120
+ "type": "ACTION",
121
+ "shape": [
122
+ 8
123
+ ]
124
+ },
125
+ "action.skill_id": {
126
+ "type": "ACTION",
127
+ "shape": [
128
+ 1
129
+ ]
130
+ }
131
+ },
132
+ "device": "cuda",
133
+ "use_amp": false,
134
+ "use_peft": false,
135
+ "push_to_hub": true,
136
+ "repo_id": null,
137
+ "private": null,
138
+ "tags": null,
139
+ "license": null,
140
+ "pretrained_path": null,
141
+ "chunk_size": 16,
142
+ "n_action_steps": 16,
143
+ "max_state_dim": 64,
144
+ "max_action_dim": 32,
145
+ "normalization_mapping": {
146
+ "VISUAL": "IDENTITY",
147
+ "STATE": "MEAN_STD",
148
+ "ACTION": "MEAN_STD"
149
+ },
150
+ "image_size": [
151
+ 224,
152
+ 224
153
+ ],
154
+ "base_model_path": "nvidia/GR00T-N1.5-3B",
155
+ "tokenizer_assets_repo": "lerobot/eagle2hg-processor-groot-n1p5",
156
+ "embodiment_tag": "new_embodiment",
157
+ "tune_llm": false,
158
+ "tune_visual": false,
159
+ "tune_projector": true,
160
+ "tune_diffusion_model": true,
161
+ "lora_rank": 0,
162
+ "lora_alpha": 16,
163
+ "lora_dropout": 0.1,
164
+ "lora_full_model": false,
165
+ "optimizer_lr": 0.0001,
166
+ "optimizer_betas": [
167
+ 0.95,
168
+ 0.999
169
+ ],
170
+ "optimizer_eps": 1e-08,
171
+ "optimizer_weight_decay": 1e-05,
172
+ "warmup_ratio": 0.05,
173
+ "use_bf16": true,
174
+ "video_backend": "pyav",
175
+ "balance_dataset_weights": true,
176
+ "balance_trajectory_weights": true,
177
+ "dataset_paths": null,
178
+ "output_dir": "./tmp/gr00t",
179
+ "save_steps": 1000,
180
+ "max_steps": 10000,
181
+ "batch_size": 32,
182
+ "dataloader_num_workers": 8,
183
+ "report_to": "wandb",
184
+ "resume": false
185
+ },
186
+ "output_dir": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
187
+ "job_name": "groot_insight",
188
+ "resume": false,
189
+ "seed": 42,
190
+ "cudnn_deterministic": false,
191
+ "num_workers": 8,
192
+ "batch_size": 64,
193
+ "steps": 100000,
194
+ "eval_freq": 20000,
195
+ "log_freq": 100,
196
+ "tolerance_s": 0.0001,
197
+ "save_checkpoint": true,
198
+ "save_freq": 50000,
199
+ "use_policy_training_preset": true,
200
+ "optimizer": null,
201
+ "scheduler": null,
202
+ "eval": {
203
+ "n_episodes": 50,
204
+ "batch_size": 50,
205
+ "use_async_envs": false,
206
+ "max_episodes_rendered": 10
207
+ },
208
+ "wandb": {
209
+ "enable": true,
210
+ "disable_artifact": false,
211
+ "project": "groot_insight_baseline_base",
212
+ "entity": "RwHlabs",
213
+ "notes": null,
214
+ "run_id": null,
215
+ "mode": null,
216
+ "add_tags": true,
217
+ "name": "insight_phase1_dit_only_groot",
218
+ "group": "insight_phase1_dit_only",
219
+ "tags": "base,phase2,groot,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
220
+ },
221
+ "peft": null,
222
+ "use_rabc": false,
223
+ "rabc_progress_path": null,
224
+ "rabc_kappa": 0.01,
225
+ "rabc_epsilon": 1e-06,
226
+ "rabc_head_mode": "sparse",
227
+ "rename_map": {},
228
+ "checkpoint_path": null,
229
+ "method": "base",
230
+ "phase": "phase2",
231
+ "pretrained_path": "",
232
+ "gradient_checkpointing": false,
233
+ "smoke_test": false,
234
+ "dry_run": false,
235
+ "grad_clip_norm": 10.0,
236
+ "write_reports": true,
237
+ "fail_on_language_fallback": true,
238
+ "action_dim": 8,
239
+ "state_dim": 16,
240
+ "camera_set": "rgb",
241
+ "video_feature_keys": "",
242
+ "task_prompt_mode": "non_guide",
243
+ "task_descriptions_path": "",
244
+ "excluded_indices_path": "",
245
+ "exclude_index_scope": "none",
246
+ "recipe_yaml": "",
247
+ "pipeline_yaml": "",
248
+ "pipeline_mode": "both",
249
+ "dry_print_commands": false,
250
+ "launcher": "python",
251
+ "phase2_output_dir": "",
252
+ "phase3_output_dir": "",
253
+ "wandb_auto_name": true,
254
+ "run_name": "",
255
+ "wandb_project_prefix": "groot_insight",
256
+ "wandb_group": "",
257
+ "wandb_tags": ""
258
+ }
checkpoints/100000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,335 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 0.0,
4
+ "betas": [
5
+ 0.95,
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
+ 159,
179
+ 160,
180
+ 161,
181
+ 162,
182
+ 163,
183
+ 164,
184
+ 165,
185
+ 166,
186
+ 167,
187
+ 168,
188
+ 169,
189
+ 170,
190
+ 171,
191
+ 172,
192
+ 173,
193
+ 174,
194
+ 175,
195
+ 176,
196
+ 177,
197
+ 178,
198
+ 179,
199
+ 180,
200
+ 181,
201
+ 182,
202
+ 183,
203
+ 184,
204
+ 185,
205
+ 186,
206
+ 187,
207
+ 188,
208
+ 189,
209
+ 190,
210
+ 191,
211
+ 192,
212
+ 193,
213
+ 194,
214
+ 195,
215
+ 196,
216
+ 197,
217
+ 198,
218
+ 199,
219
+ 200,
220
+ 201,
221
+ 202,
222
+ 203,
223
+ 204,
224
+ 205,
225
+ 206,
226
+ 207,
227
+ 208,
228
+ 209,
229
+ 210,
230
+ 211,
231
+ 212,
232
+ 213,
233
+ 214,
234
+ 215,
235
+ 216,
236
+ 217,
237
+ 218,
238
+ 219,
239
+ 220,
240
+ 221,
241
+ 222,
242
+ 223,
243
+ 224,
244
+ 225,
245
+ 226,
246
+ 227,
247
+ 228,
248
+ 229,
249
+ 230,
250
+ 231,
251
+ 232,
252
+ 233,
253
+ 234,
254
+ 235,
255
+ 236,
256
+ 237,
257
+ 238,
258
+ 239,
259
+ 240,
260
+ 241,
261
+ 242,
262
+ 243,
263
+ 244,
264
+ 245,
265
+ 246,
266
+ 247,
267
+ 248,
268
+ 249,
269
+ 250,
270
+ 251,
271
+ 252,
272
+ 253,
273
+ 254,
274
+ 255,
275
+ 256,
276
+ 257,
277
+ 258,
278
+ 259,
279
+ 260,
280
+ 261,
281
+ 262,
282
+ 263,
283
+ 264,
284
+ 265,
285
+ 266,
286
+ 267,
287
+ 268,
288
+ 269,
289
+ 270,
290
+ 271,
291
+ 272,
292
+ 273,
293
+ 274,
294
+ 275,
295
+ 276,
296
+ 277,
297
+ 278,
298
+ 279,
299
+ 280,
300
+ 281,
301
+ 282,
302
+ 283,
303
+ 284,
304
+ 285,
305
+ 286,
306
+ 287,
307
+ 288,
308
+ 289,
309
+ 290,
310
+ 291,
311
+ 292,
312
+ 293,
313
+ 294,
314
+ 295,
315
+ 296,
316
+ 297,
317
+ 298,
318
+ 299,
319
+ 300,
320
+ 301,
321
+ 302,
322
+ 303,
323
+ 304,
324
+ 305,
325
+ 306,
326
+ 307,
327
+ 308,
328
+ 309,
329
+ 310,
330
+ 311,
331
+ 312,
332
+ 313
333
+ ]
334
+ }
335
+ ]
checkpoints/100000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:538fd88805669011d3587f7dbce7727d1ba801e4649a970275978e5b963057ff
3
+ size 4275306416
checkpoints/100000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34b815630d4ce3f560a5b0a3d07e08fd8cd8a03c15fdd6df6ef15a81e44ceb7f
3
+ size 15708
checkpoints/100000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 0.0001
4
+ ],
5
+ "last_epoch": 100000,
6
+ "_step_count": 100001,
7
+ "_get_lr_called_within_step": false,
8
+ "_last_lr": [
9
+ 0.0
10
+ ],
11
+ "lr_lambdas": [
12
+ {}
13
+ ]
14
+ }
checkpoints/100000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 100000
3
+ }
dataset_sample_report.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_shape": [
3
+ 16,
4
+ 8
5
+ ],
6
+ "camera_keys_present": {
7
+ "observation.images.guide": true,
8
+ "observation.images.right_shoulder": true,
9
+ "observation.images.wrist": true
10
+ },
11
+ "camera_order_effective": [
12
+ "observation.images.guide",
13
+ "observation.images.right_shoulder",
14
+ "observation.images.wrist"
15
+ ],
16
+ "keys_after_filter_slice": [
17
+ "action",
18
+ "action.skill_id",
19
+ "action_is_pad",
20
+ "episode_index",
21
+ "frame_index",
22
+ "index",
23
+ "next.reward",
24
+ "observation.images.guide",
25
+ "observation.images.right_shoulder",
26
+ "observation.images.wrist",
27
+ "observation.state",
28
+ "task",
29
+ "task_index",
30
+ "timestamp"
31
+ ],
32
+ "observation_state_shape": [
33
+ 16
34
+ ],
35
+ "sample_index": 0,
36
+ "task": "1ext",
37
+ "task_index": {
38
+ "preview": [
39
+ 0
40
+ ],
41
+ "shape": []
42
+ }
43
+ }
first_batch_report.json ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim_effective": 8,
3
+ "batch_spec": {
4
+ "action_effective_dim": 8,
5
+ "action_horizon": 16,
6
+ "action_order": [
7
+ "action"
8
+ ],
9
+ "action_storage_dim": 32,
10
+ "camera_order_effective": [
11
+ "observation.images.guide",
12
+ "observation.images.right_shoulder",
13
+ "observation.images.wrist"
14
+ ],
15
+ "chunk_size": 16,
16
+ "padded_action_dim": 32,
17
+ "padded_state_dim": 64,
18
+ "state_effective_dim": 16,
19
+ "state_order": [
20
+ "observation.state"
21
+ ],
22
+ "state_storage_dim": 32
23
+ },
24
+ "camera_order_effective": [
25
+ "observation.images.guide",
26
+ "observation.images.right_shoulder",
27
+ "observation.images.wrist"
28
+ ],
29
+ "camera_set": "rgb",
30
+ "dataset_len": 246793,
31
+ "language_fallback_detected": false,
32
+ "language_fallback_scan": {
33
+ "all_paths": [],
34
+ "detected": false,
35
+ "processed_paths": [],
36
+ "raw_paths": []
37
+ },
38
+ "method": "base",
39
+ "phase": "phase2",
40
+ "policy_type": "groot",
41
+ "processed_action_shape": [
42
+ 64,
43
+ 16,
44
+ 32
45
+ ],
46
+ "processed_state_shape": [
47
+ 64,
48
+ 1,
49
+ 64
50
+ ],
51
+ "raw_action_shape": [
52
+ 64,
53
+ 16,
54
+ 8
55
+ ],
56
+ "raw_keys": [
57
+ "action",
58
+ "action.skill_id",
59
+ "action_is_pad",
60
+ "episode_index",
61
+ "frame_index",
62
+ "index",
63
+ "next.reward",
64
+ "observation.images.guide",
65
+ "observation.images.right_shoulder",
66
+ "observation.images.wrist",
67
+ "observation.state",
68
+ "task",
69
+ "task_index",
70
+ "timestamp"
71
+ ],
72
+ "raw_observation_state_shape": [
73
+ 64,
74
+ 16
75
+ ],
76
+ "raw_task_preview": [
77
+ "Open the bottle following the guide.",
78
+ "Open the bottle following the guide.",
79
+ "Open the door following the guide."
80
+ ],
81
+ "script_name": "train_groot_insight_yaml.py",
82
+ "state_dim_effective": 16,
83
+ "task_prompt_mode": "non_guide",
84
+ "train_batch_tensor_report": {
85
+ "action": {
86
+ "device": "cuda:0",
87
+ "dtype": "torch.float32",
88
+ "shape": [
89
+ 64,
90
+ 16,
91
+ 32
92
+ ]
93
+ },
94
+ "action_is_pad": {
95
+ "device": "cuda:0",
96
+ "dtype": "torch.bool",
97
+ "shape": [
98
+ 64,
99
+ 16
100
+ ]
101
+ },
102
+ "action_mask": {
103
+ "device": "cuda:0",
104
+ "dtype": "torch.bool",
105
+ "shape": [
106
+ 64,
107
+ 16,
108
+ 32
109
+ ]
110
+ },
111
+ "eagle_attention_mask": {
112
+ "device": "cuda:0",
113
+ "dtype": "torch.int64",
114
+ "shape": [
115
+ 64,
116
+ 816
117
+ ]
118
+ },
119
+ "eagle_image_sizes": {
120
+ "device": "cuda:0",
121
+ "dtype": "torch.int64",
122
+ "shape": [
123
+ 192,
124
+ 2
125
+ ]
126
+ },
127
+ "eagle_input_ids": {
128
+ "device": "cuda:0",
129
+ "dtype": "torch.int64",
130
+ "shape": [
131
+ 64,
132
+ 816
133
+ ]
134
+ },
135
+ "eagle_pixel_values": {
136
+ "device": "cuda:0",
137
+ "dtype": "torch.float32",
138
+ "shape": [
139
+ 192,
140
+ 3,
141
+ 224,
142
+ 224
143
+ ]
144
+ },
145
+ "embodiment_id": {
146
+ "device": "cuda:0",
147
+ "dtype": "torch.int64",
148
+ "shape": [
149
+ 64
150
+ ]
151
+ },
152
+ "episode_index": {
153
+ "device": "cuda:0",
154
+ "dtype": "torch.int64",
155
+ "shape": [
156
+ 64
157
+ ]
158
+ },
159
+ "index": {
160
+ "device": "cuda:0",
161
+ "dtype": "torch.int64",
162
+ "shape": [
163
+ 64
164
+ ]
165
+ },
166
+ "next.reward": {
167
+ "device": "cuda:0",
168
+ "dtype": "torch.float32",
169
+ "shape": [
170
+ 64
171
+ ]
172
+ },
173
+ "observation.state": {
174
+ "device": "cuda:0",
175
+ "dtype": "torch.float32",
176
+ "shape": [
177
+ 64,
178
+ 16
179
+ ]
180
+ },
181
+ "state": {
182
+ "device": "cuda:0",
183
+ "dtype": "torch.float32",
184
+ "shape": [
185
+ 64,
186
+ 1,
187
+ 64
188
+ ]
189
+ },
190
+ "state_mask": {
191
+ "device": "cuda:0",
192
+ "dtype": "torch.bool",
193
+ "shape": [
194
+ 64,
195
+ 1,
196
+ 64
197
+ ]
198
+ },
199
+ "task_index": {
200
+ "device": "cuda:0",
201
+ "dtype": "torch.int64",
202
+ "shape": [
203
+ 64
204
+ ]
205
+ }
206
+ }
207
+ }
pipeline_commands.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "phase2_args": {
3
+ "batch_size": 64,
4
+ "camera_set": "rgb",
5
+ "dataset.repo_id": "/home/seonho/insight/insight_dataset",
6
+ "dataset.root": "/home/seonho/insight/insight_dataset",
7
+ "dataset.video_backend": "pyav",
8
+ "log_freq": 100,
9
+ "num_workers": 8,
10
+ "output_dir": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
11
+ "phase": "phase2",
12
+ "policy.base_model_path": "nvidia/GR00T-N1.5-3B",
13
+ "policy.lora_rank": 0,
14
+ "policy.tune_diffusion_model": true,
15
+ "policy.tune_llm": false,
16
+ "policy.tune_projector": true,
17
+ "policy.tune_visual": false,
18
+ "policy.type": "groot",
19
+ "save_freq": 50000,
20
+ "steps": 100000,
21
+ "task_prompt_mode": "non_guide",
22
+ "wandb.enable": true,
23
+ "wandb.entity": "RwHlabs",
24
+ "wandb.group": "insight_phase1_dit_only",
25
+ "wandb.name": "insight_phase1_dit_only_groot",
26
+ "wandb.project": "groot_insight_baseline_base",
27
+ "wandb.tags": "base,phase2,groot,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
28
+ },
29
+ "phase2_command": [
30
+ "accelerate",
31
+ "launch",
32
+ "--num_processes",
33
+ "1",
34
+ "--mixed_precision",
35
+ "no",
36
+ "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
37
+ "--batch_size=64",
38
+ "--camera_set=rgb",
39
+ "--dataset.repo_id=/home/seonho/insight/insight_dataset",
40
+ "--dataset.root=/home/seonho/insight/insight_dataset",
41
+ "--dataset.video_backend=pyav",
42
+ "--log_freq=100",
43
+ "--num_workers=8",
44
+ "--output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt",
45
+ "--phase=phase2",
46
+ "--policy.base_model_path=nvidia/GR00T-N1.5-3B",
47
+ "--policy.lora_rank=0",
48
+ "--policy.tune_diffusion_model=true",
49
+ "--policy.tune_llm=false",
50
+ "--policy.tune_projector=true",
51
+ "--policy.tune_visual=false",
52
+ "--policy.type=groot",
53
+ "--save_freq=50000",
54
+ "--steps=100000",
55
+ "--task_prompt_mode=non_guide",
56
+ "--wandb.enable=true",
57
+ "--wandb.entity=RwHlabs",
58
+ "--wandb.group=insight_phase1_dit_only",
59
+ "--wandb.name=insight_phase1_dit_only_groot",
60
+ "--wandb.project=groot_insight_baseline_base",
61
+ "--wandb.tags=base,phase2,groot,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
62
+ ],
63
+ "phase3_args": null,
64
+ "phase3_command": null
65
+ }
pipeline_manifest.json ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "end_timestamp": "2026-05-23T05:32:47.031197+00:00",
3
+ "git_commit": "f8482718aaa75ce2ad60b9b8dd7070a607da4249",
4
+ "launcher": "accelerate launch --num_processes 1 --mixed_precision no",
5
+ "phase2_command": [
6
+ "accelerate",
7
+ "launch",
8
+ "--num_processes",
9
+ "1",
10
+ "--mixed_precision",
11
+ "no",
12
+ "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
13
+ "--batch_size=64",
14
+ "--camera_set=rgb",
15
+ "--dataset.repo_id=/home/seonho/insight/insight_dataset",
16
+ "--dataset.root=/home/seonho/insight/insight_dataset",
17
+ "--dataset.video_backend=pyav",
18
+ "--log_freq=100",
19
+ "--num_workers=8",
20
+ "--output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt",
21
+ "--phase=phase2",
22
+ "--policy.base_model_path=nvidia/GR00T-N1.5-3B",
23
+ "--policy.lora_rank=0",
24
+ "--policy.tune_diffusion_model=true",
25
+ "--policy.tune_llm=false",
26
+ "--policy.tune_projector=true",
27
+ "--policy.tune_visual=false",
28
+ "--policy.type=groot",
29
+ "--save_freq=50000",
30
+ "--steps=100000",
31
+ "--task_prompt_mode=non_guide",
32
+ "--wandb.enable=true",
33
+ "--wandb.entity=RwHlabs",
34
+ "--wandb.group=insight_phase1_dit_only",
35
+ "--wandb.name=insight_phase1_dit_only_groot",
36
+ "--wandb.project=groot_insight_baseline_base",
37
+ "--wandb.tags=base,phase2,groot,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
38
+ ],
39
+ "phase2_output_dir": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
40
+ "phase3_command": null,
41
+ "phase3_output_dir": null,
42
+ "pipeline_mode": "phase2_only",
43
+ "recipe_yaml": "/home/seonho/clvla/Experiments_cfg/insight/phase1_dit_only.yaml",
44
+ "resolved_phase2_checkpoint": null,
45
+ "return_status": "success",
46
+ "start_timestamp": "2026-05-19T13:13:01.780734+00:00",
47
+ "wandb_resolved": {
48
+ "group": "insight_phase1_dit_only",
49
+ "phase2": {
50
+ "enable": true,
51
+ "entity": "RwHlabs",
52
+ "group": "insight_phase1_dit_only",
53
+ "method": "base",
54
+ "name": "insight_phase1_dit_only_groot",
55
+ "phase": "phase2",
56
+ "project": "groot_insight_baseline_base",
57
+ "tags": [
58
+ "base",
59
+ "phase2",
60
+ "groot",
61
+ "rgb",
62
+ "non_guide",
63
+ "insight",
64
+ "insight_native",
65
+ "cartesian_8d",
66
+ "phase1",
67
+ "baseline",
68
+ "dit_only",
69
+ "nvidia_base"
70
+ ]
71
+ },
72
+ "phase2_name": "insight_phase1_dit_only_groot",
73
+ "phase3": null,
74
+ "phase3_name": null,
75
+ "project": "groot_insight_baseline_base",
76
+ "tags": [
77
+ "base",
78
+ "phase2",
79
+ "groot",
80
+ "rgb",
81
+ "non_guide",
82
+ "insight",
83
+ "insight_native",
84
+ "cartesian_8d",
85
+ "phase1",
86
+ "baseline",
87
+ "dit_only",
88
+ "nvidia_base"
89
+ ]
90
+ }
91
+ }
run_manifest.json ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_dim_effective": 8,
3
+ "action_dim_storage": 32,
4
+ "batch_spec": {
5
+ "action_effective_dim": 8,
6
+ "action_horizon": 16,
7
+ "action_order": [
8
+ "action"
9
+ ],
10
+ "action_storage_dim": 32,
11
+ "camera_order_effective": [
12
+ "observation.images.guide",
13
+ "observation.images.right_shoulder",
14
+ "observation.images.wrist"
15
+ ],
16
+ "chunk_size": 16,
17
+ "padded_action_dim": 32,
18
+ "padded_state_dim": 64,
19
+ "state_effective_dim": 16,
20
+ "state_order": [
21
+ "observation.state"
22
+ ],
23
+ "state_storage_dim": 32
24
+ },
25
+ "camera_order_effective": [
26
+ "observation.images.guide",
27
+ "observation.images.right_shoulder",
28
+ "observation.images.wrist"
29
+ ],
30
+ "camera_set": "rgb",
31
+ "checkpoint_processor_state_loaded": false,
32
+ "checkpoint_processor_state_note": "phase3 loads model.safetensors weights only; pre/post processors are rebuilt from the current INSIGHT dataset stats for normalization consistency.",
33
+ "cl_v2_fm_loss_weight": null,
34
+ "cl_v2_loss_weight": null,
35
+ "cl_v2_phase": null,
36
+ "cl_v2_trainable_mode": null,
37
+ "created_at_utc": "2026-05-19T13:13:32.874286+00:00",
38
+ "dataset_len": 246793,
39
+ "dataset_root": "/home/seonho/insight/insight_dataset",
40
+ "exclude_index_scope": "none",
41
+ "excluded_indices_path": "",
42
+ "fps": 10,
43
+ "loaded_pretrained_model_path": null,
44
+ "method": "base",
45
+ "mgd_enabled": null,
46
+ "mgd_loss_weight": null,
47
+ "mgd_trainable_mode": null,
48
+ "phase": "phase2",
49
+ "policy_type": "groot",
50
+ "preprocessor_stats_source": "INSIGHT dataset stats sliced to effective action/state dims",
51
+ "pretrained_path": "",
52
+ "repo_id": "/home/seonho/insight/insight_dataset",
53
+ "rkd_loss_weight": null,
54
+ "script_name": "train_groot_insight_yaml.py",
55
+ "state_dim_effective": 16,
56
+ "state_dim_storage": 32,
57
+ "task_prompt_mode": "non_guide",
58
+ "total_episodes_metadata": 5647,
59
+ "total_frames_metadata": 246793,
60
+ "total_tasks_metadata": 13,
61
+ "trainable_report": {
62
+ "dit_core_only_unexpected_trainable_params": 0,
63
+ "groups": {
64
+ "_groot_model.action_head.model": {
65
+ "total": 550386688,
66
+ "trainable": 550386688
67
+ },
68
+ "_groot_model.action_head.vl_self_attention": {
69
+ "total": 201433088,
70
+ "trainable": 201433088
71
+ },
72
+ "_groot_model.action_head.vlln": {
73
+ "total": 4096,
74
+ "trainable": 4096
75
+ },
76
+ "sequence_mgd_head": {
77
+ "total": 0,
78
+ "trainable": 0
79
+ }
80
+ },
81
+ "num_learnable_params": 1068806144,
82
+ "num_total_params": 2724163520,
83
+ "other_trainable_params": 1068806144,
84
+ "policy_type": "groot",
85
+ "selected_prefixes": [],
86
+ "trainable_mode": null,
87
+ "trainable_ratio": 0.3923428737493702
88
+ },
89
+ "wandb_resolved": {
90
+ "camera_set": "rgb",
91
+ "enable": true,
92
+ "entity": "RwHlabs",
93
+ "group": "insight_phase1_dit_only",
94
+ "method": "base",
95
+ "name": "insight_phase1_dit_only_groot",
96
+ "phase": "phase2",
97
+ "project": "groot_insight_baseline_base",
98
+ "tags": [
99
+ "base",
100
+ "phase2",
101
+ "groot",
102
+ "rgb",
103
+ "non_guide",
104
+ "insight",
105
+ "insight_native",
106
+ "cartesian_8d",
107
+ "phase1",
108
+ "baseline",
109
+ "dit_only",
110
+ "nvidia_base"
111
+ ],
112
+ "task_prompt_mode": "non_guide"
113
+ }
114
+ }
train_summary_report.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "final_checkpoint": "/home/seonho/insight/outputs/dit_only_phase1_ckpt/checkpoints/100000",
3
+ "last_grad_norm": 1.2890625,
4
+ "last_loss_finite": true,
5
+ "last_lr": 0.0,
6
+ "last_policy_update_called": false,
7
+ "last_scalars": {
8
+ "loss": 0.06392817199230194
9
+ },
10
+ "loaded_pretrained_model_path": null,
11
+ "method": "base",
12
+ "phase": "phase2",
13
+ "policy_type": "groot",
14
+ "script_name": "train_groot_insight_yaml.py",
15
+ "steps_completed": 100000,
16
+ "steps_requested": 100000
17
+ }
trainable_param_report.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dit_core_only_unexpected_trainable_params": 0,
3
+ "groups": {
4
+ "_groot_model.action_head.model": {
5
+ "total": 550386688,
6
+ "trainable": 550386688
7
+ },
8
+ "_groot_model.action_head.vl_self_attention": {
9
+ "total": 201433088,
10
+ "trainable": 201433088
11
+ },
12
+ "_groot_model.action_head.vlln": {
13
+ "total": 4096,
14
+ "trainable": 4096
15
+ },
16
+ "sequence_mgd_head": {
17
+ "total": 0,
18
+ "trainable": 0
19
+ }
20
+ },
21
+ "num_learnable_params": 1068806144,
22
+ "num_total_params": 2724163520,
23
+ "other_trainable_params": 1068806144,
24
+ "policy_type": "groot",
25
+ "selected_prefixes": [],
26
+ "trainable_mode": null,
27
+ "trainable_ratio": 0.3923428737493702
28
+ }
wandb/debug-internal.log ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-05-19T22:13:33.757508612+09:00","level":"INFO","msg":"stream: starting","core version":"0.24.2"}
2
+ {"time":"2026-05-19T22:13:34.249681253+09:00","level":"INFO","msg":"stream: created new stream","id":"qz452p60"}
3
+ {"time":"2026-05-19T22:13:34.249816267+09:00","level":"INFO","msg":"handler: started","stream_id":"qz452p60"}
4
+ {"time":"2026-05-19T22:13:34.250122028+09:00","level":"INFO","msg":"stream: started","id":"qz452p60"}
5
+ {"time":"2026-05-19T22:13:34.250159759+09:00","level":"INFO","msg":"sender: started","stream_id":"qz452p60"}
6
+ {"time":"2026-05-19T22:13:34.250156101+09:00","level":"INFO","msg":"writer: started","stream_id":"qz452p60"}
7
+ {"time":"2026-05-20T15:55:10.818156537+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
8
+ {"time":"2026-05-20T15:56:44.994984532+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
9
+ {"time":"2026-05-20T15:58:23.494572793+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
10
+ {"time":"2026-05-20T17:32:56.207023413+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
11
+ {"time":"2026-05-20T17:33:53.844718799+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
12
+ {"time":"2026-05-20T17:34:53.908902815+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
13
+ {"time":"2026-05-20T17:35:57.323694482+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
14
+ {"time":"2026-05-20T18:30:49.563054911+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
15
+ {"time":"2026-05-20T18:48:36.484690792+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
16
+ {"time":"2026-05-20T18:49:38.236616601+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
17
+ {"time":"2026-05-20T18:51:49.732015936+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
18
+ {"time":"2026-05-20T18:59:43.774277496+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
19
+ {"time":"2026-05-20T19:03:49.908408418+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
20
+ {"time":"2026-05-20T19:04:48.543201774+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
21
+ {"time":"2026-05-20T19:05:45.963581069+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
22
+ {"time":"2026-05-20T19:07:50.284610418+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
23
+ {"time":"2026-05-20T19:12:30.299479148+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
24
+ {"time":"2026-05-20T19:14:13.564300748+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
25
+ {"time":"2026-05-20T19:15:14.101427763+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
26
+ {"time":"2026-05-20T19:16:15.40790845+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
27
+ {"time":"2026-05-20T19:29:32.060104195+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
28
+ {"time":"2026-05-20T19:31:30.297052013+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
29
+ {"time":"2026-05-20T19:49:26.161299534+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
30
+ {"time":"2026-05-20T19:50:23.803656458+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
31
+ {"time":"2026-05-20T20:04:29.31970088+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
32
+ {"time":"2026-05-20T20:49:42.028043576+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
33
+ {"time":"2026-05-20T20:50:44.339074594+09:00","level":"INFO","msg":"api: retrying HTTP error","status":500,"url":"https://api.wandb.ai/files/RwHlabs/groot_insight_baseline_base/qz452p60/file_stream","body":"{\"error\":\"context deadline exceeded\"}"}
34
+ {"time":"2026-05-23T13:56:03.488569868+09:00","level":"INFO","msg":"api: retrying error","error":"unexpected HTTP status 502 Bad Gateway"}
35
+ {"time":"2026-05-23T14:25:43.511326685+09:00","level":"INFO","msg":"api: retrying error","error":"unexpected HTTP status 502 Bad Gateway"}
36
+ {"time":"2026-05-23T14:32:39.967322696+09:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
37
+ {"time":"2026-05-23T14:32:40.260656129+09:00","level":"INFO","msg":"handler: operation stats","stats":{}}
38
+ {"time":"2026-05-23T14:32:40.266561008+09:00","level":"INFO","msg":"stream: closing","id":"qz452p60"}
39
+ {"time":"2026-05-23T14:32:40.266576703+09:00","level":"INFO","msg":"handler: closed","stream_id":"qz452p60"}
40
+ {"time":"2026-05-23T14:32:40.266689508+09:00","level":"INFO","msg":"sender: closed","stream_id":"qz452p60"}
41
+ {"time":"2026-05-23T14:32:40.266726133+09:00","level":"INFO","msg":"stream: closed","id":"qz452p60"}
wandb/debug.log ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_setup.py:_flush():81] Current SDK version is 0.24.2
2
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_setup.py:_flush():81] Configure stats pid to 3941155
3
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /home/seonho/insight/outputs/dit_only_phase1_ckpt/wandb/run-20260519_221333-qz452p60/logs/debug.log
5
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /home/seonho/insight/outputs/dit_only_phase1_ckpt/wandb/run-20260519_221333-qz452p60/logs/debug-internal.log
6
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_init.py:init():844] calling init triggers
7
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
8
+ config: {'repo_id': '/home/seonho/insight/insight_dataset', 'dataset_root': '/home/seonho/insight/insight_dataset', 'method': 'base', 'phase': 'phase2', 'steps': 100000, 'lr': 0.0001, 'batch_size': 64, 'num_processes': 1, 'effective_batch_size': 64, 'action_dim': 8, 'state_dim': 16, 'camera_set': 'rgb', 'video_feature_keys': '', 'task_prompt_mode': 'non_guide', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_target': None, 'mgd_enabled': None, 'mgd_trainable_mode': None, 'mgd_loss_weight': None, 'mgd_backprop_backbone': None, 'cl_v2_phase': None, 'cl_v2_trainable_mode': None, 'cl_v2_loss_weight': None, 'cl_v2_fm_loss_weight': None, 'gradient_checkpointing': False, 'seed': 42, 'dataset_len': 246793, '_wandb': {}}
9
+ 2026-05-19 22:13:33,530 INFO MainThread:3941155 [wandb_init.py:init():892] starting backend
10
+ 2026-05-19 22:13:33,746 INFO MainThread:3941155 [wandb_init.py:init():895] sending inform_init request
11
+ 2026-05-19 22:13:33,753 INFO MainThread:3941155 [wandb_init.py:init():903] backend started and connected
12
+ 2026-05-19 22:13:33,755 INFO MainThread:3941155 [wandb_init.py:init():973] updated telemetry
13
+ 2026-05-19 22:13:33,756 INFO MainThread:3941155 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
14
+ 2026-05-19 22:13:35,113 INFO MainThread:3941155 [wandb_init.py:init():1042] starting run threads in backend
15
+ 2026-05-19 22:13:35,241 INFO MainThread:3941155 [wandb_run.py:_console_start():2529] atexit reg
16
+ 2026-05-19 22:13:35,241 INFO MainThread:3941155 [wandb_run.py:_redirect():2377] redirect: wrap_raw
17
+ 2026-05-19 22:13:35,241 INFO MainThread:3941155 [wandb_run.py:_redirect():2446] Wrapping output streams.
18
+ 2026-05-19 22:13:35,241 INFO MainThread:3941155 [wandb_run.py:_redirect():2469] Redirects installed.
19
+ 2026-05-19 22:13:35,246 INFO MainThread:3941155 [wandb_init.py:init():1082] run started, returning control to user process
20
+ 2026-05-23 13:53:31,185 INFO MainThread:3941155 [wandb_run.py:_finish():2295] finishing run RwHlabs/groot_insight_baseline_base/qz452p60
21
+ 2026-05-23 13:53:31,187 INFO MainThread:3941155 [wandb_run.py:_atexit_cleanup():2494] got exitcode: 0
22
+ 2026-05-23 13:53:31,187 INFO MainThread:3941155 [wandb_run.py:_restore():2476] restore
23
+ 2026-05-23 13:53:31,187 INFO MainThread:3941155 [wandb_run.py:_restore():2482] restore done
24
+ 2026-05-23 14:32:40,265 INFO MainThread:3941155 [wandb_run.py:_footer_sync_info():3871] logging synced files
wandb/run-20260518_152847-smy9mkg4/files/config.yaml ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ _wandb:
2
+ value:
3
+ cli_version: 0.24.2
4
+ e:
5
+ ys7pyc71ohdki3edek8hpsuxl4fkwwj9:
6
+ args:
7
+ - --action_dim=8
8
+ - --batch_size=64
9
+ - --camera_set=rgb
10
+ - --dataset.repo_id=/home/seonho/insight/insight_dataset
11
+ - --dataset.root=/home/seonho/insight/insight_dataset
12
+ - --dataset.video_backend=torchcodec
13
+ - --log_freq=100
14
+ - --num_workers=4
15
+ - --output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt
16
+ - --phase=phase2
17
+ - --policy.base_model_path=nvidia/GR00T-N1.5-3B
18
+ - --policy.lora_rank=0
19
+ - --policy.mgd_enabled=false
20
+ - --policy.mgd_loss_weight=0.0
21
+ - --policy.mgd_trainable_mode=dit_only
22
+ - --policy.tune_llm=false
23
+ - --policy.tune_visual=false
24
+ - --policy.type=groot_processed_mgd
25
+ - --save_freq=50000
26
+ - --state_dim=16
27
+ - --steps=100000
28
+ - --task_prompt_mode=non_guide
29
+ - --wandb.enable=true
30
+ - --wandb.entity=RwHlabs
31
+ - --wandb.group=insight_phase1_dit_only
32
+ - --wandb.name=insight_phase1_dit_only_groot
33
+ - --wandb.project=groot_insight_baseline_mgd
34
+ - --wandb.tags=mgd,phase2,groot_processed_mgd,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base
35
+ codePath: clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
36
+ codePathLocal: clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
37
+ cpu_count: 32
38
+ cpu_count_logical: 64
39
+ cudaVersion: "12.2"
40
+ disk:
41
+ /:
42
+ total: "3778130485248"
43
+ used: "3526731341824"
44
+ email: dankk98@gmail.com
45
+ executable: /home/seonho/miniconda3/envs/minje/bin/python3.12
46
+ gpu: NVIDIA A100 80GB PCIe
47
+ gpu_count: 4
48
+ gpu_nvidia:
49
+ - architecture: Ampere
50
+ cudaCores: 6912
51
+ memoryTotal: "85899345920"
52
+ name: NVIDIA A100 80GB PCIe
53
+ uuid: GPU-0dd6d491-2dd8-24ac-bb27-872483ab77e9
54
+ - architecture: Ampere
55
+ cudaCores: 6912
56
+ memoryTotal: "85899345920"
57
+ name: NVIDIA A100 80GB PCIe
58
+ uuid: GPU-dbfb7925-aec1-50fb-039a-3a4c0a4b2caa
59
+ - architecture: Ampere
60
+ cudaCores: 6912
61
+ memoryTotal: "85899345920"
62
+ name: NVIDIA A100 80GB PCIe
63
+ uuid: GPU-ca1c0319-e178-a3c3-ac41-386c2723deac
64
+ - architecture: Ampere
65
+ cudaCores: 6912
66
+ memoryTotal: "85899345920"
67
+ name: NVIDIA A100 80GB PCIe
68
+ uuid: GPU-47b6036e-4fde-e399-5515-f73eed752c6d
69
+ host: worker1
70
+ memory:
71
+ total: "1081620840448"
72
+ os: Linux-6.8.0-101-generic-x86_64-with-glibc2.35
73
+ program: /home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
74
+ python: CPython 3.12.12
75
+ root: /home/seonho/insight/outputs/dit_only_phase1_ckpt
76
+ startedAt: "2026-05-18T06:28:47.096325Z"
77
+ writerId: ys7pyc71ohdki3edek8hpsuxl4fkwwj9
78
+ m: []
79
+ python_version: 3.12.12
80
+ t:
81
+ "1":
82
+ - 1
83
+ - 5
84
+ - 11
85
+ - 41
86
+ - 49
87
+ - 51
88
+ - 53
89
+ - 71
90
+ - 83
91
+ - 98
92
+ "2":
93
+ - 1
94
+ - 5
95
+ - 11
96
+ - 41
97
+ - 49
98
+ - 51
99
+ - 53
100
+ - 71
101
+ - 83
102
+ - 98
103
+ "3":
104
+ - 13
105
+ - 15
106
+ - 16
107
+ - 62
108
+ "4": 3.12.12
109
+ "5": 0.24.2
110
+ "6": 5.3.0
111
+ "12": 0.24.2
112
+ "13": linux-x86_64
113
+ action_dim:
114
+ value: 8
115
+ batch_size:
116
+ value: 64
117
+ camera_set:
118
+ value: rgb
119
+ cl_v2_fm_loss_weight:
120
+ value: null
121
+ cl_v2_loss_weight:
122
+ value: null
123
+ cl_v2_phase:
124
+ value: null
125
+ cl_v2_trainable_mode:
126
+ value: null
127
+ dataset_len:
128
+ value: 246793
129
+ dataset_root:
130
+ value: /home/seonho/insight/insight_dataset
131
+ effective_batch_size:
132
+ value: 64
133
+ gradient_checkpointing:
134
+ value: false
135
+ lora_alpha:
136
+ value: 16
137
+ lora_dropout:
138
+ value: 0.1
139
+ lora_rank:
140
+ value: 0
141
+ lora_target:
142
+ value: llm
143
+ lr:
144
+ value: 0.0001
145
+ method:
146
+ value: mgd
147
+ mgd_backprop_backbone:
148
+ value: true
149
+ mgd_enabled:
150
+ value: false
151
+ mgd_loss_weight:
152
+ value: 0
153
+ mgd_trainable_mode:
154
+ value: dit_only
155
+ num_processes:
156
+ value: 1
157
+ phase:
158
+ value: phase2
159
+ repo_id:
160
+ value: /home/seonho/insight/insight_dataset
161
+ seed:
162
+ value: 42
163
+ state_dim:
164
+ value: 16
165
+ steps:
166
+ value: 100000
167
+ task_prompt_mode:
168
+ value: non_guide
169
+ tune_diffusion_model:
170
+ value: true
171
+ tune_llm:
172
+ value: false
173
+ tune_projector:
174
+ value: true
175
+ tune_visual:
176
+ value: false
177
+ video_feature_keys:
178
+ value: ""
wandb/run-20260518_152847-smy9mkg4/files/output.log ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Traceback (most recent call last):
2
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1735, in <module>
3
+ main()
4
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/configs/parser.py", line 233, in wrapper_inner
5
+ response = fn(cfg, *args, **kwargs)
6
+ ^^^^^^^^^^^^^^^^^^^^^^^^
7
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1621, in main
8
+ raw_batch = prepare_raw_batch(next(data_stream), cfg, preset, task_desc_map)
9
+ ^^^^^^^^^^^^^^^^^
10
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1453, in infinite_dataloader
11
+ yield from dataloader
12
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/accelerate/data_loader.py", line 577, in __iter__
13
+ current_batch = next(dataloader_iter)
14
+ ^^^^^^^^^^^^^^^^^^^^^
15
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 733, in __next__
16
+ data = self._next_data()
17
+ ^^^^^^^^^^^^^^^^^
18
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 1515, in _next_data
19
+ return self._process_data(data, worker_id)
20
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
21
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 1550, in _process_data
22
+ data.reraise()
23
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/_utils.py", line 750, in reraise
24
+ raise exception
25
+ RuntimeError: Caught RuntimeError in DataLoader worker process 0.
26
+ Original Traceback (most recent call last):
27
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/_utils/worker.py", line 349, in _worker_loop
28
+ data = fetcher.fetch(index) # type: ignore[possibly-undefined]
29
+ ^^^^^^^^^^^^^^^^^^^^
30
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
31
+ data = [self.dataset[idx] for idx in possibly_batched_index]
32
+ ~~~~~~~~~~~~^^^^^
33
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/lerobot_dataset.py", line 627, in __getitem__
34
+ video_frames = self._query_videos(query_timestamps, ep_idx)
35
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
36
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/lerobot_dataset.py", line 590, in _query_videos
37
+ frames = decode_video_frames(video_path, shifted_query_ts, self.tolerance_s, self.video_backend)
38
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
39
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/video_utils.py", line 152, in decode_video_frames
40
+ return decode_video_frames_torchcodec(video_path, timestamps, tolerance_s)
41
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
42
+ File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/video_utils.py", line 420, in decode_video_frames_torchcodec
43
+ frames_batch = decoder.get_frames_at(indices=frame_indices)
44
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
45
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torchcodec/decoders/_video_decoder.py", line 227, in get_frames_at
46
+ data, pts_seconds, duration_seconds = core.get_frames_at_indices(
47
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^
48
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/_ops.py", line 756, in __call__
49
+ return self._op(*args, **kwargs)
50
+ ^^^^^^^^^^^^^^^^^^^^^^^^^
51
+ RuntimeError: Could not push packet to decoder: Invalid data found when processing input
52
+
53
+ [rank0]: Traceback (most recent call last):
54
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1735, in <module>
55
+ [rank0]: main()
56
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/configs/parser.py", line 233, in wrapper_inner
57
+ [rank0]: response = fn(cfg, *args, **kwargs)
58
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^
59
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1621, in main
60
+ [rank0]: raw_batch = prepare_raw_batch(next(data_stream), cfg, preset, task_desc_map)
61
+ [rank0]: ^^^^^^^^^^^^^^^^^
62
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py", line 1453, in infinite_dataloader
63
+ [rank0]: yield from dataloader
64
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/accelerate/data_loader.py", line 577, in __iter__
65
+ [rank0]: current_batch = next(dataloader_iter)
66
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^
67
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 733, in __next__
68
+ [rank0]: data = self._next_data()
69
+ [rank0]: ^^^^^^^^^^^^^^^^^
70
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 1515, in _next_data
71
+ [rank0]: return self._process_data(data, worker_id)
72
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
73
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/dataloader.py", line 1550, in _process_data
74
+ [rank0]: data.reraise()
75
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/_utils.py", line 750, in reraise
76
+ [rank0]: raise exception
77
+ [rank0]: RuntimeError: Caught RuntimeError in DataLoader worker process 0.
78
+ [rank0]: Original Traceback (most recent call last):
79
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/_utils/worker.py", line 349, in _worker_loop
80
+ [rank0]: data = fetcher.fetch(index) # type: ignore[possibly-undefined]
81
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^
82
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
83
+ [rank0]: data = [self.dataset[idx] for idx in possibly_batched_index]
84
+ [rank0]: ~~~~~~~~~~~~^^^^^
85
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/lerobot_dataset.py", line 627, in __getitem__
86
+ [rank0]: video_frames = self._query_videos(query_timestamps, ep_idx)
87
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
88
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/lerobot_dataset.py", line 590, in _query_videos
89
+ [rank0]: frames = decode_video_frames(video_path, shifted_query_ts, self.tolerance_s, self.video_backend)
90
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
91
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/video_utils.py", line 152, in decode_video_frames
92
+ [rank0]: return decode_video_frames_torchcodec(video_path, timestamps, tolerance_s)
93
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
94
+ [rank0]: File "/home/seonho/clvla/lerobot_cl/src/lerobot/datasets/video_utils.py", line 420, in decode_video_frames_torchcodec
95
+ [rank0]: frames_batch = decoder.get_frames_at(indices=frame_indices)
96
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
97
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torchcodec/decoders/_video_decoder.py", line 227, in get_frames_at
98
+ [rank0]: data, pts_seconds, duration_seconds = core.get_frames_at_indices(
99
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^
100
+ [rank0]: File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/torch/_ops.py", line 756, in __call__
101
+ [rank0]: return self._op(*args, **kwargs)
102
+ [rank0]: ^^^^^^^^^^^^^^^^^^^^^^^^^
103
+ [rank0]: RuntimeError: Could not push packet to decoder: Invalid data found when processing input
wandb/run-20260518_152847-smy9mkg4/files/requirements.txt ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ attrs==25.4.0
2
+ kiwisolver==1.5.0
3
+ setuptools==80.10.2
4
+ transformers==5.3.0
5
+ pynndescent==0.6.0
6
+ filelock==3.25.2
7
+ numba==0.61.2
8
+ gymnasium==1.2.3
9
+ shellingham==1.5.4
10
+ dill==0.3.8
11
+ ninja==1.13.0
12
+ hidapi==0.15.0
13
+ scikit-image==0.26.0
14
+ orderly-set==5.5.0
15
+ mypy_extensions==1.1.0
16
+ stringzilla==4.6.1
17
+ tianshou==0.4.10
18
+ numpy==2.2.5
19
+ xxhash==3.6.0
20
+ protobuf==3.19.6
21
+ numpydantic==1.6.7
22
+ pyserial==3.5
23
+ omegaconf==2.3.0
24
+ Farama-Notifications==0.0.4
25
+ nvidia-nccl-cu12==2.26.2
26
+ plotly==6.6.0
27
+ platformdirs==4.9.4
28
+ nvidia-cufft-cu12==11.3.0.4
29
+ eval_type_backport==0.3.1
30
+ packaging==25.0
31
+ typing-inspect==0.9.0
32
+ opencv-python-headless==4.12.0.88
33
+ cmeel-assimp==6.0.2
34
+ lerobot==0.5.1
35
+ pyyaml-include==1.4.1
36
+ urllib3==2.6.3
37
+ lxml==6.1.0
38
+ peft==0.18.1
39
+ virtualenv==21.2.4
40
+ cmeel==0.59.0
41
+ pre_commit==4.6.0
42
+ importlib_metadata==8.7.1
43
+ click==8.3.1
44
+ accelerate==1.13.0
45
+ multiprocess==0.70.16
46
+ distlib==0.4.0
47
+ websockets==16.0
48
+ albumentations==1.4.18
49
+ tokenizers==0.22.2
50
+ nvidia-cusparselt-cu12==0.6.3
51
+ certifi==2026.2.25
52
+ annotated-doc==0.0.4
53
+ triton==3.3.1
54
+ libpinocchio==3.9.0
55
+ mpmath==1.3.0
56
+ decord==0.6.0
57
+ matplotlib==3.10.8
58
+ llvmlite==0.44.0
59
+ python-xlib==0.33
60
+ regex==2026.2.28
61
+ colored==2.3.2
62
+ charset-normalizer==3.4.5
63
+ mdurl==0.1.2
64
+ GitPython==3.1.46
65
+ gym==0.26.2
66
+ av==15.1.0
67
+ markdown-it-py==4.0.0
68
+ glfw==2.10.0
69
+ nvidia-cufile-cu12==1.11.1.6
70
+ PyYAML==6.0.3
71
+ python-dateutil==2.9.0.post0
72
+ cmeel-tinyxml2==10.0.0
73
+ blinker==1.9.0
74
+ opencv-python==4.13.0.92
75
+ cmake==4.1.3
76
+ gym-notices==0.1.0
77
+ typing-inspection==0.4.2
78
+ timm==1.0.25
79
+ torchcodec==0.5
80
+ cmeel-octomap==1.10.0
81
+ umap-learn==0.5.11
82
+ MarkupSafe==3.0.3
83
+ pluggy==1.6.0
84
+ smmap==5.0.3
85
+ fsspec==2025.3.0
86
+ evdev==1.9.3
87
+ python-discovery==1.2.2
88
+ hf-xet==1.5.0
89
+ scikit-learn==1.8.0
90
+ h5py==3.16.0
91
+ cmeel-console-bridge==1.0.2.3
92
+ narwhals==2.18.1
93
+ msgpack==1.1.2
94
+ flash_attn==2.8.3
95
+ ansitable==0.11.5
96
+ pyarrow==23.0.1
97
+ h11==0.16.0
98
+ mergedeep==1.3.4
99
+ tqdm==4.67.3
100
+ vla-eval==0.1.1.dev27+g195d46338
101
+ nvidia-nvjitlink-cu12==12.6.85
102
+ cmeel-qhull==8.0.2.1
103
+ libcoal==3.0.2
104
+ cmeel-boost==1.89.0
105
+ rtb-data==1.0.1
106
+ mujoco==3.3.1
107
+ nvidia-cuda-nvrtc-cu12==12.6.77
108
+ PyOpenGL==3.1.10
109
+ draccus==0.10.0
110
+ pip==26.0.1
111
+ threadpoolctl==3.6.0
112
+ annotated-types==0.7.0
113
+ anyio==4.12.1
114
+ imageio-ffmpeg==0.6.0
115
+ qpsolvers==4.11.0
116
+ wrapt==2.1.2
117
+ networkx==3.6.1
118
+ contourpy==1.3.3
119
+ Werkzeug==3.1.8
120
+ tensorboard-data-server==0.7.2
121
+ gitdb==4.0.12
122
+ nvidia-cusparse-cu12==12.5.4.2
123
+ typer==0.24.1
124
+ toml==0.10.2
125
+ tensorboard==2.20.0
126
+ rerun-sdk==0.26.2
127
+ diffusers==0.35.2
128
+ albucore==0.0.17
129
+ wandb==0.24.2
130
+ nvidia-cublas-cu12==12.6.4.1
131
+ nvidia-cuda-runtime-cu12==12.6.77
132
+ cycler==0.12.1
133
+ coal==3.0.2
134
+ pydantic_core==2.27.2
135
+ torch==2.7.1
136
+ nvidia-cuda-cupti-cu12==12.6.80
137
+ Pygments==2.19.2
138
+ pytest==9.0.3
139
+ ImageIO==2.37.3
140
+ sentry-sdk==2.54.0
141
+ antlr4-python3-runtime==4.9.3
142
+ sympy==1.14.0
143
+ safetensors==0.7.0
144
+ cmeel-zlib==1.3.1
145
+ einops==0.8.2
146
+ spatialgeometry==1.1.0
147
+ pygame==2.6.1
148
+ absl-py==2.4.0
149
+ pandas==3.0.1
150
+ httpx==0.28.1
151
+ robosuite==1.5.2
152
+ fonttools==4.62.1
153
+ progress==1.6.1
154
+ nvidia-cusolver-cu12==11.7.1.2
155
+ aiohttp==3.13.3
156
+ frozenlist==1.8.0
157
+ deepdiff==8.6.1
158
+ yarl==1.23.0
159
+ nvidia-cudnn-cu12==9.5.1.17
160
+ torchvision==0.22.1
161
+ idna==3.11
162
+ lazy-loader==0.5
163
+ termcolor==3.3.0
164
+ nvidia-curand-cu12==10.3.7.77
165
+ multidict==6.7.1
166
+ Markdown==3.10.2
167
+ pynput==1.8.1
168
+ propcache==0.4.1
169
+ zipp==3.23.0
170
+ nvidia-nvtx-cu12==12.6.77
171
+ Flask==3.1.3
172
+ httpcore==1.0.9
173
+ etils==1.14.0
174
+ psutil==7.2.2
175
+ pillow==12.1.1
176
+ nodeenv==1.10.0
177
+ pin==3.9.0
178
+ quadprog==0.1.13
179
+ eigenpy==3.12.0
180
+ wheel==0.46.3
181
+ simsimd==6.5.16
182
+ itsdangerous==2.2.0
183
+ aiohappyeyeballs==2.6.1
184
+ cfgv==3.5.0
185
+ cloudpickle==3.1.2
186
+ iniconfig==2.3.0
187
+ requests==2.32.5
188
+ tifffile==2026.5.2
189
+ joblib==1.5.3
190
+ identify==2.6.19
191
+ huggingface_hub==1.14.0
192
+ datasets==4.8.4
193
+ typing_extensions==4.15.0
194
+ pydantic==2.10.6
195
+ jsonlines==4.0.0
196
+ swift-sim==1.1.0
197
+ lazyregistry==0.4.0
198
+ robocasa==1.0.0
199
+ pgraph-python==0.6.3
200
+ Jinja2==3.1.6
201
+ aiosignal==1.4.0
202
+ dm-tree==0.1.9
203
+ six==1.17.0
204
+ grpcio==1.80.0
205
+ rich==14.3.3
206
+ scipy==1.15.3
207
+ cmeel-urdfdom==4.0.1
208
+ pyparsing==3.3.2
209
+ lerobot==0.5.1
wandb/run-20260518_152847-smy9mkg4/files/wandb-metadata.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-6.8.0-101-generic-x86_64-with-glibc2.35",
3
+ "python": "CPython 3.12.12",
4
+ "startedAt": "2026-05-18T06:28:47.096325Z",
5
+ "args": [
6
+ "--action_dim=8",
7
+ "--batch_size=64",
8
+ "--camera_set=rgb",
9
+ "--dataset.repo_id=/home/seonho/insight/insight_dataset",
10
+ "--dataset.root=/home/seonho/insight/insight_dataset",
11
+ "--dataset.video_backend=torchcodec",
12
+ "--log_freq=100",
13
+ "--num_workers=4",
14
+ "--output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt",
15
+ "--phase=phase2",
16
+ "--policy.base_model_path=nvidia/GR00T-N1.5-3B",
17
+ "--policy.lora_rank=0",
18
+ "--policy.mgd_enabled=false",
19
+ "--policy.mgd_loss_weight=0.0",
20
+ "--policy.mgd_trainable_mode=dit_only",
21
+ "--policy.tune_llm=false",
22
+ "--policy.tune_visual=false",
23
+ "--policy.type=groot_processed_mgd",
24
+ "--save_freq=50000",
25
+ "--state_dim=16",
26
+ "--steps=100000",
27
+ "--task_prompt_mode=non_guide",
28
+ "--wandb.enable=true",
29
+ "--wandb.entity=RwHlabs",
30
+ "--wandb.group=insight_phase1_dit_only",
31
+ "--wandb.name=insight_phase1_dit_only_groot",
32
+ "--wandb.project=groot_insight_baseline_mgd",
33
+ "--wandb.tags=mgd,phase2,groot_processed_mgd,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
34
+ ],
35
+ "program": "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
36
+ "codePath": "clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
37
+ "codePathLocal": "clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
38
+ "email": "dankk98@gmail.com",
39
+ "root": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
40
+ "host": "worker1",
41
+ "executable": "/home/seonho/miniconda3/envs/minje/bin/python3.12",
42
+ "cpu_count": 32,
43
+ "cpu_count_logical": 64,
44
+ "gpu": "NVIDIA A100 80GB PCIe",
45
+ "gpu_count": 4,
46
+ "disk": {
47
+ "/": {
48
+ "total": "3778130485248",
49
+ "used": "3526731341824"
50
+ }
51
+ },
52
+ "memory": {
53
+ "total": "1081620840448"
54
+ },
55
+ "gpu_nvidia": [
56
+ {
57
+ "name": "NVIDIA A100 80GB PCIe",
58
+ "memoryTotal": "85899345920",
59
+ "cudaCores": 6912,
60
+ "architecture": "Ampere",
61
+ "uuid": "GPU-0dd6d491-2dd8-24ac-bb27-872483ab77e9"
62
+ },
63
+ {
64
+ "name": "NVIDIA A100 80GB PCIe",
65
+ "memoryTotal": "85899345920",
66
+ "cudaCores": 6912,
67
+ "architecture": "Ampere",
68
+ "uuid": "GPU-dbfb7925-aec1-50fb-039a-3a4c0a4b2caa"
69
+ },
70
+ {
71
+ "name": "NVIDIA A100 80GB PCIe",
72
+ "memoryTotal": "85899345920",
73
+ "cudaCores": 6912,
74
+ "architecture": "Ampere",
75
+ "uuid": "GPU-ca1c0319-e178-a3c3-ac41-386c2723deac"
76
+ },
77
+ {
78
+ "name": "NVIDIA A100 80GB PCIe",
79
+ "memoryTotal": "85899345920",
80
+ "cudaCores": 6912,
81
+ "architecture": "Ampere",
82
+ "uuid": "GPU-47b6036e-4fde-e399-5515-f73eed752c6d"
83
+ }
84
+ ],
85
+ "cudaVersion": "12.2",
86
+ "writerId": "ys7pyc71ohdki3edek8hpsuxl4fkwwj9"
87
+ }
wandb/run-20260518_152847-smy9mkg4/files/wandb-summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"params/whole_total":2427681217,"params/whole_trainable":840593408,"_wandb":{"runtime":1},"_runtime":1}
wandb/run-20260518_152847-smy9mkg4/logs/debug-core.log ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-05-18T15:28:47.134214528+09:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmp6v9hfjeh/port-3560821.txt","pid":3560821,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
2
+ {"time":"2026-05-18T15:28:47.134775705+09:00","level":"INFO","msg":"server: will exit if parent process dies","ppid":3560821}
3
+ {"time":"2026-05-18T15:28:47.134766733+09:00","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-3560821-3566957-720346231/socket","Net":"unix"}}
4
+ {"time":"2026-05-18T15:28:47.322025048+09:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
5
+ {"time":"2026-05-18T15:28:47.336551071+09:00","level":"INFO","msg":"handleInformInit: received","streamId":"smy9mkg4","id":"1(@)"}
6
+ {"time":"2026-05-18T15:28:47.715349144+09:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"smy9mkg4","id":"1(@)"}
7
+ {"time":"2026-05-18T15:28:50.380217573+09:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
8
+ {"time":"2026-05-18T15:28:50.380327768+09:00","level":"INFO","msg":"server is shutting down"}
9
+ {"time":"2026-05-18T15:28:50.380307348+09:00","level":"INFO","msg":"connection: closing","id":"1(@)"}
10
+ {"time":"2026-05-18T15:28:50.380576767+09:00","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
11
+ {"time":"2026-05-18T15:28:50.38058255+09:00","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-3560821-3566957-720346231/socket","Net":"unix"}}
12
+ {"time":"2026-05-18T15:28:52.389072306+09:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
13
+ {"time":"2026-05-18T15:28:52.389105339+09:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
14
+ {"time":"2026-05-18T15:28:52.389126226+09:00","level":"INFO","msg":"server is closed"}
wandb/run-20260518_152847-smy9mkg4/logs/debug-internal.log ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-05-18T15:28:47.336692013+09:00","level":"INFO","msg":"stream: starting","core version":"0.24.2"}
2
+ {"time":"2026-05-18T15:28:47.71499011+09:00","level":"INFO","msg":"stream: created new stream","id":"smy9mkg4"}
3
+ {"time":"2026-05-18T15:28:47.715138854+09:00","level":"INFO","msg":"handler: started","stream_id":"smy9mkg4"}
4
+ {"time":"2026-05-18T15:28:47.715332098+09:00","level":"INFO","msg":"stream: started","id":"smy9mkg4"}
5
+ {"time":"2026-05-18T15:28:47.715342118+09:00","level":"INFO","msg":"writer: started","stream_id":"smy9mkg4"}
6
+ {"time":"2026-05-18T15:28:47.715343845+09:00","level":"INFO","msg":"sender: started","stream_id":"smy9mkg4"}
7
+ {"time":"2026-05-18T15:28:50.380317028+09:00","level":"INFO","msg":"stream: closing","id":"smy9mkg4"}
8
+ {"time":"2026-05-18T15:28:52.036130686+09:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
9
+ {"time":"2026-05-18T15:28:52.388015098+09:00","level":"INFO","msg":"handler: closed","stream_id":"smy9mkg4"}
10
+ {"time":"2026-05-18T15:28:52.388154428+09:00","level":"INFO","msg":"sender: closed","stream_id":"smy9mkg4"}
11
+ {"time":"2026-05-18T15:28:52.388166789+09:00","level":"INFO","msg":"stream: closed","id":"smy9mkg4"}
wandb/run-20260518_152847-smy9mkg4/logs/debug.log ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_setup.py:_flush():81] Current SDK version is 0.24.2
2
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_setup.py:_flush():81] Configure stats pid to 3560821
3
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /home/seonho/insight/outputs/dit_only_phase1_ckpt/wandb/run-20260518_152847-smy9mkg4/logs/debug.log
5
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /home/seonho/insight/outputs/dit_only_phase1_ckpt/wandb/run-20260518_152847-smy9mkg4/logs/debug-internal.log
6
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_init.py:init():844] calling init triggers
7
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
8
+ config: {'repo_id': '/home/seonho/insight/insight_dataset', 'dataset_root': '/home/seonho/insight/insight_dataset', 'method': 'mgd', 'phase': 'phase2', 'steps': 100000, 'lr': 0.0001, 'batch_size': 64, 'num_processes': 1, 'effective_batch_size': 64, 'action_dim': 8, 'state_dim': 16, 'camera_set': 'rgb', 'video_feature_keys': '', 'task_prompt_mode': 'non_guide', 'tune_llm': False, 'tune_visual': False, 'tune_projector': True, 'tune_diffusion_model': True, 'lora_rank': 0, 'lora_alpha': 16, 'lora_dropout': 0.1, 'lora_target': 'llm', 'mgd_enabled': False, 'mgd_trainable_mode': 'dit_only', 'mgd_loss_weight': 0.0, 'mgd_backprop_backbone': True, 'cl_v2_phase': None, 'cl_v2_trainable_mode': None, 'cl_v2_loss_weight': None, 'cl_v2_fm_loss_weight': None, 'gradient_checkpointing': False, 'seed': 42, 'dataset_len': 246793, '_wandb': {}}
9
+ 2026-05-18 15:28:47,098 INFO MainThread:3560821 [wandb_init.py:init():892] starting backend
10
+ 2026-05-18 15:28:47,322 INFO MainThread:3560821 [wandb_init.py:init():895] sending inform_init request
11
+ 2026-05-18 15:28:47,332 INFO MainThread:3560821 [wandb_init.py:init():903] backend started and connected
12
+ 2026-05-18 15:28:47,335 INFO MainThread:3560821 [wandb_init.py:init():973] updated telemetry
13
+ 2026-05-18 15:28:47,336 INFO MainThread:3560821 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
14
+ 2026-05-18 15:28:48,978 INFO MainThread:3560821 [wandb_init.py:init():1042] starting run threads in backend
15
+ 2026-05-18 15:28:49,171 INFO MainThread:3560821 [wandb_run.py:_console_start():2529] atexit reg
16
+ 2026-05-18 15:28:49,171 INFO MainThread:3560821 [wandb_run.py:_redirect():2377] redirect: wrap_raw
17
+ 2026-05-18 15:28:49,171 INFO MainThread:3560821 [wandb_run.py:_redirect():2446] Wrapping output streams.
18
+ 2026-05-18 15:28:49,171 INFO MainThread:3560821 [wandb_run.py:_redirect():2469] Redirects installed.
19
+ 2026-05-18 15:28:49,176 INFO MainThread:3560821 [wandb_init.py:init():1082] run started, returning control to user process
20
+ 2026-05-18 15:28:50,380 INFO wandb-AsyncioManager-main:3560821 [service_client.py:_forward_responses():94] Reached EOF.
21
+ 2026-05-18 15:28:50,381 INFO wandb-AsyncioManager-main:3560821 [mailbox.py:close():154] Closing mailbox, abandoning 1 handles.
22
+ 2026-05-18 15:28:52,576 ERROR wandb-AsyncioManager-main:3560821 [asyncio_manager.py:fn_wrap_exceptions():183] Uncaught exception in run_soon callback.
23
+ Traceback (most recent call last):
24
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/wandb/sdk/lib/asyncio_manager.py", line 181, in fn_wrap_exceptions
25
+ await fn()
26
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 38, in publish
27
+ await self._send_server_request(request)
28
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/site-packages/wandb/sdk/lib/service/service_client.py", line 64, in _send_server_request
29
+ await self._writer.drain()
30
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/asyncio/streams.py", line 392, in drain
31
+ await self._protocol._drain_helper()
32
+ File "/home/seonho/miniconda3/envs/minje/lib/python3.12/asyncio/streams.py", line 166, in _drain_helper
33
+ raise ConnectionResetError('Connection lost')
34
+ ConnectionResetError: Connection lost
wandb/run-20260518_152847-smy9mkg4/run-smy9mkg4.wandb ADDED
Binary file (12.9 kB). View file
 
wandb/run-20260518_153314-yn3lgfkj/files/config.yaml ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ _wandb:
2
+ value:
3
+ cli_version: 0.24.2
4
+ e:
5
+ ddeaytz4ga2t8xmz3osihmmadhyuvdau:
6
+ args:
7
+ - --action_dim=8
8
+ - --batch_size=64
9
+ - --camera_set=rgb
10
+ - --dataset.repo_id=/home/seonho/insight/insight_dataset
11
+ - --dataset.root=/home/seonho/insight/insight_dataset
12
+ - --dataset.video_backend=pyav
13
+ - --log_freq=100
14
+ - --num_workers=4
15
+ - --output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt
16
+ - --phase=phase2
17
+ - --policy.base_model_path=nvidia/GR00T-N1.5-3B
18
+ - --policy.lora_rank=0
19
+ - --policy.mgd_enabled=false
20
+ - --policy.mgd_loss_weight=0.0
21
+ - --policy.mgd_trainable_mode=dit_only
22
+ - --policy.tune_llm=false
23
+ - --policy.tune_visual=false
24
+ - --policy.type=groot_processed_mgd
25
+ - --save_freq=50000
26
+ - --state_dim=16
27
+ - --steps=100000
28
+ - --task_prompt_mode=non_guide
29
+ - --wandb.enable=true
30
+ - --wandb.entity=RwHlabs
31
+ - --wandb.group=insight_phase1_dit_only
32
+ - --wandb.name=insight_phase1_dit_only_groot
33
+ - --wandb.project=groot_insight_baseline_mgd
34
+ - --wandb.tags=mgd,phase2,groot_processed_mgd,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base
35
+ codePath: clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
36
+ codePathLocal: clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
37
+ cpu_count: 32
38
+ cpu_count_logical: 64
39
+ cudaVersion: "12.2"
40
+ disk:
41
+ /:
42
+ total: "3778130485248"
43
+ used: "3442052386816"
44
+ email: dankk98@gmail.com
45
+ executable: /home/seonho/miniconda3/envs/minje/bin/python3.12
46
+ gpu: NVIDIA A100 80GB PCIe
47
+ gpu_count: 4
48
+ gpu_nvidia:
49
+ - architecture: Ampere
50
+ cudaCores: 6912
51
+ memoryTotal: "85899345920"
52
+ name: NVIDIA A100 80GB PCIe
53
+ uuid: GPU-0dd6d491-2dd8-24ac-bb27-872483ab77e9
54
+ - architecture: Ampere
55
+ cudaCores: 6912
56
+ memoryTotal: "85899345920"
57
+ name: NVIDIA A100 80GB PCIe
58
+ uuid: GPU-dbfb7925-aec1-50fb-039a-3a4c0a4b2caa
59
+ - architecture: Ampere
60
+ cudaCores: 6912
61
+ memoryTotal: "85899345920"
62
+ name: NVIDIA A100 80GB PCIe
63
+ uuid: GPU-ca1c0319-e178-a3c3-ac41-386c2723deac
64
+ - architecture: Ampere
65
+ cudaCores: 6912
66
+ memoryTotal: "85899345920"
67
+ name: NVIDIA A100 80GB PCIe
68
+ uuid: GPU-47b6036e-4fde-e399-5515-f73eed752c6d
69
+ host: worker1
70
+ memory:
71
+ total: "1081620840448"
72
+ os: Linux-6.8.0-101-generic-x86_64-with-glibc2.35
73
+ program: /home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py
74
+ python: CPython 3.12.12
75
+ root: /home/seonho/insight/outputs/dit_only_phase1_ckpt
76
+ startedAt: "2026-05-18T06:33:14.327419Z"
77
+ writerId: ddeaytz4ga2t8xmz3osihmmadhyuvdau
78
+ m: []
79
+ python_version: 3.12.12
80
+ t:
81
+ "1":
82
+ - 1
83
+ - 5
84
+ - 11
85
+ - 41
86
+ - 49
87
+ - 51
88
+ - 53
89
+ - 71
90
+ - 83
91
+ - 98
92
+ "2":
93
+ - 1
94
+ - 5
95
+ - 11
96
+ - 41
97
+ - 49
98
+ - 51
99
+ - 53
100
+ - 71
101
+ - 83
102
+ - 98
103
+ "3":
104
+ - 13
105
+ - 15
106
+ - 16
107
+ - 62
108
+ "4": 3.12.12
109
+ "5": 0.24.2
110
+ "6": 5.3.0
111
+ "12": 0.24.2
112
+ "13": linux-x86_64
113
+ action_dim:
114
+ value: 8
115
+ batch_size:
116
+ value: 64
117
+ camera_set:
118
+ value: rgb
119
+ cl_v2_fm_loss_weight:
120
+ value: null
121
+ cl_v2_loss_weight:
122
+ value: null
123
+ cl_v2_phase:
124
+ value: null
125
+ cl_v2_trainable_mode:
126
+ value: null
127
+ dataset_len:
128
+ value: 246793
129
+ dataset_root:
130
+ value: /home/seonho/insight/insight_dataset
131
+ effective_batch_size:
132
+ value: 64
133
+ gradient_checkpointing:
134
+ value: false
135
+ lora_alpha:
136
+ value: 16
137
+ lora_dropout:
138
+ value: 0.1
139
+ lora_rank:
140
+ value: 0
141
+ lora_target:
142
+ value: llm
143
+ lr:
144
+ value: 0.0001
145
+ method:
146
+ value: mgd
147
+ mgd_backprop_backbone:
148
+ value: true
149
+ mgd_enabled:
150
+ value: false
151
+ mgd_loss_weight:
152
+ value: 0
153
+ mgd_trainable_mode:
154
+ value: dit_only
155
+ num_processes:
156
+ value: 1
157
+ phase:
158
+ value: phase2
159
+ repo_id:
160
+ value: /home/seonho/insight/insight_dataset
161
+ seed:
162
+ value: 42
163
+ state_dim:
164
+ value: 16
165
+ steps:
166
+ value: 100000
167
+ task_prompt_mode:
168
+ value: non_guide
169
+ tune_diffusion_model:
170
+ value: true
171
+ tune_llm:
172
+ value: false
173
+ tune_projector:
174
+ value: true
175
+ tune_visual:
176
+ value: false
177
+ video_feature_keys:
178
+ value: ""
wandb/run-20260518_153314-yn3lgfkj/files/output.log ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ `use_fast` is set to `True` but the image processor class does not have a fast version. Falling back to the slow version.
2
+ The tokenizer you are loading from '/home/seonho/.cache/huggingface/lerobot/lerobot/eagle2hg-processor-groot-n1p5' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
3
+ The tokenizer you are loading from '/home/seonho/.cache/huggingface/lerobot/lerobot/eagle2hg-processor-groot-n1p5' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
wandb/run-20260518_153314-yn3lgfkj/files/requirements.txt ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ attrs==25.4.0
2
+ kiwisolver==1.5.0
3
+ setuptools==80.10.2
4
+ transformers==5.3.0
5
+ pynndescent==0.6.0
6
+ filelock==3.25.2
7
+ numba==0.61.2
8
+ gymnasium==1.2.3
9
+ shellingham==1.5.4
10
+ dill==0.3.8
11
+ ninja==1.13.0
12
+ hidapi==0.15.0
13
+ scikit-image==0.26.0
14
+ orderly-set==5.5.0
15
+ mypy_extensions==1.1.0
16
+ stringzilla==4.6.1
17
+ tianshou==0.4.10
18
+ numpy==2.2.5
19
+ xxhash==3.6.0
20
+ protobuf==3.19.6
21
+ numpydantic==1.6.7
22
+ pyserial==3.5
23
+ omegaconf==2.3.0
24
+ Farama-Notifications==0.0.4
25
+ nvidia-nccl-cu12==2.26.2
26
+ plotly==6.6.0
27
+ platformdirs==4.9.4
28
+ nvidia-cufft-cu12==11.3.0.4
29
+ eval_type_backport==0.3.1
30
+ packaging==25.0
31
+ typing-inspect==0.9.0
32
+ opencv-python-headless==4.12.0.88
33
+ cmeel-assimp==6.0.2
34
+ lerobot==0.5.1
35
+ pyyaml-include==1.4.1
36
+ urllib3==2.6.3
37
+ lxml==6.1.0
38
+ peft==0.18.1
39
+ virtualenv==21.2.4
40
+ cmeel==0.59.0
41
+ pre_commit==4.6.0
42
+ importlib_metadata==8.7.1
43
+ click==8.3.1
44
+ accelerate==1.13.0
45
+ multiprocess==0.70.16
46
+ distlib==0.4.0
47
+ websockets==16.0
48
+ albumentations==1.4.18
49
+ tokenizers==0.22.2
50
+ nvidia-cusparselt-cu12==0.6.3
51
+ certifi==2026.2.25
52
+ annotated-doc==0.0.4
53
+ triton==3.3.1
54
+ libpinocchio==3.9.0
55
+ mpmath==1.3.0
56
+ decord==0.6.0
57
+ matplotlib==3.10.8
58
+ llvmlite==0.44.0
59
+ python-xlib==0.33
60
+ regex==2026.2.28
61
+ colored==2.3.2
62
+ charset-normalizer==3.4.5
63
+ mdurl==0.1.2
64
+ GitPython==3.1.46
65
+ gym==0.26.2
66
+ av==15.1.0
67
+ markdown-it-py==4.0.0
68
+ glfw==2.10.0
69
+ nvidia-cufile-cu12==1.11.1.6
70
+ PyYAML==6.0.3
71
+ python-dateutil==2.9.0.post0
72
+ cmeel-tinyxml2==10.0.0
73
+ blinker==1.9.0
74
+ opencv-python==4.13.0.92
75
+ cmake==4.1.3
76
+ gym-notices==0.1.0
77
+ typing-inspection==0.4.2
78
+ timm==1.0.25
79
+ torchcodec==0.5
80
+ cmeel-octomap==1.10.0
81
+ umap-learn==0.5.11
82
+ MarkupSafe==3.0.3
83
+ pluggy==1.6.0
84
+ smmap==5.0.3
85
+ fsspec==2025.3.0
86
+ evdev==1.9.3
87
+ python-discovery==1.2.2
88
+ hf-xet==1.5.0
89
+ scikit-learn==1.8.0
90
+ h5py==3.16.0
91
+ cmeel-console-bridge==1.0.2.3
92
+ narwhals==2.18.1
93
+ msgpack==1.1.2
94
+ flash_attn==2.8.3
95
+ ansitable==0.11.5
96
+ pyarrow==23.0.1
97
+ h11==0.16.0
98
+ mergedeep==1.3.4
99
+ tqdm==4.67.3
100
+ vla-eval==0.1.1.dev27+g195d46338
101
+ nvidia-nvjitlink-cu12==12.6.85
102
+ cmeel-qhull==8.0.2.1
103
+ libcoal==3.0.2
104
+ cmeel-boost==1.89.0
105
+ rtb-data==1.0.1
106
+ mujoco==3.3.1
107
+ nvidia-cuda-nvrtc-cu12==12.6.77
108
+ PyOpenGL==3.1.10
109
+ draccus==0.10.0
110
+ pip==26.0.1
111
+ threadpoolctl==3.6.0
112
+ annotated-types==0.7.0
113
+ anyio==4.12.1
114
+ imageio-ffmpeg==0.6.0
115
+ qpsolvers==4.11.0
116
+ wrapt==2.1.2
117
+ networkx==3.6.1
118
+ contourpy==1.3.3
119
+ Werkzeug==3.1.8
120
+ tensorboard-data-server==0.7.2
121
+ gitdb==4.0.12
122
+ nvidia-cusparse-cu12==12.5.4.2
123
+ typer==0.24.1
124
+ toml==0.10.2
125
+ tensorboard==2.20.0
126
+ rerun-sdk==0.26.2
127
+ diffusers==0.35.2
128
+ albucore==0.0.17
129
+ wandb==0.24.2
130
+ nvidia-cublas-cu12==12.6.4.1
131
+ nvidia-cuda-runtime-cu12==12.6.77
132
+ cycler==0.12.1
133
+ coal==3.0.2
134
+ pydantic_core==2.27.2
135
+ torch==2.7.1
136
+ nvidia-cuda-cupti-cu12==12.6.80
137
+ Pygments==2.19.2
138
+ pytest==9.0.3
139
+ ImageIO==2.37.3
140
+ sentry-sdk==2.54.0
141
+ antlr4-python3-runtime==4.9.3
142
+ sympy==1.14.0
143
+ safetensors==0.7.0
144
+ cmeel-zlib==1.3.1
145
+ einops==0.8.2
146
+ spatialgeometry==1.1.0
147
+ pygame==2.6.1
148
+ absl-py==2.4.0
149
+ pandas==3.0.1
150
+ httpx==0.28.1
151
+ robosuite==1.5.2
152
+ fonttools==4.62.1
153
+ progress==1.6.1
154
+ nvidia-cusolver-cu12==11.7.1.2
155
+ aiohttp==3.13.3
156
+ frozenlist==1.8.0
157
+ deepdiff==8.6.1
158
+ yarl==1.23.0
159
+ nvidia-cudnn-cu12==9.5.1.17
160
+ torchvision==0.22.1
161
+ idna==3.11
162
+ lazy-loader==0.5
163
+ termcolor==3.3.0
164
+ nvidia-curand-cu12==10.3.7.77
165
+ multidict==6.7.1
166
+ Markdown==3.10.2
167
+ pynput==1.8.1
168
+ propcache==0.4.1
169
+ zipp==3.23.0
170
+ nvidia-nvtx-cu12==12.6.77
171
+ Flask==3.1.3
172
+ httpcore==1.0.9
173
+ etils==1.14.0
174
+ psutil==7.2.2
175
+ pillow==12.1.1
176
+ nodeenv==1.10.0
177
+ pin==3.9.0
178
+ quadprog==0.1.13
179
+ eigenpy==3.12.0
180
+ wheel==0.46.3
181
+ simsimd==6.5.16
182
+ itsdangerous==2.2.0
183
+ aiohappyeyeballs==2.6.1
184
+ cfgv==3.5.0
185
+ cloudpickle==3.1.2
186
+ iniconfig==2.3.0
187
+ requests==2.32.5
188
+ tifffile==2026.5.2
189
+ joblib==1.5.3
190
+ identify==2.6.19
191
+ huggingface_hub==1.14.0
192
+ datasets==4.8.4
193
+ typing_extensions==4.15.0
194
+ pydantic==2.10.6
195
+ jsonlines==4.0.0
196
+ swift-sim==1.1.0
197
+ lazyregistry==0.4.0
198
+ robocasa==1.0.0
199
+ pgraph-python==0.6.3
200
+ Jinja2==3.1.6
201
+ aiosignal==1.4.0
202
+ dm-tree==0.1.9
203
+ six==1.17.0
204
+ grpcio==1.80.0
205
+ rich==14.3.3
206
+ scipy==1.15.3
207
+ cmeel-urdfdom==4.0.1
208
+ pyparsing==3.3.2
209
+ lerobot==0.5.1
wandb/run-20260518_153314-yn3lgfkj/files/wandb-metadata.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-6.8.0-101-generic-x86_64-with-glibc2.35",
3
+ "python": "CPython 3.12.12",
4
+ "startedAt": "2026-05-18T06:33:14.327419Z",
5
+ "args": [
6
+ "--action_dim=8",
7
+ "--batch_size=64",
8
+ "--camera_set=rgb",
9
+ "--dataset.repo_id=/home/seonho/insight/insight_dataset",
10
+ "--dataset.root=/home/seonho/insight/insight_dataset",
11
+ "--dataset.video_backend=pyav",
12
+ "--log_freq=100",
13
+ "--num_workers=4",
14
+ "--output_dir=/home/seonho/insight/outputs/dit_only_phase1_ckpt",
15
+ "--phase=phase2",
16
+ "--policy.base_model_path=nvidia/GR00T-N1.5-3B",
17
+ "--policy.lora_rank=0",
18
+ "--policy.mgd_enabled=false",
19
+ "--policy.mgd_loss_weight=0.0",
20
+ "--policy.mgd_trainable_mode=dit_only",
21
+ "--policy.tune_llm=false",
22
+ "--policy.tune_visual=false",
23
+ "--policy.type=groot_processed_mgd",
24
+ "--save_freq=50000",
25
+ "--state_dim=16",
26
+ "--steps=100000",
27
+ "--task_prompt_mode=non_guide",
28
+ "--wandb.enable=true",
29
+ "--wandb.entity=RwHlabs",
30
+ "--wandb.group=insight_phase1_dit_only",
31
+ "--wandb.name=insight_phase1_dit_only_groot",
32
+ "--wandb.project=groot_insight_baseline_mgd",
33
+ "--wandb.tags=mgd,phase2,groot_processed_mgd,rgb,non_guide,insight,insight_native,cartesian_8d,phase1,baseline,dit_only,nvidia_base"
34
+ ],
35
+ "program": "/home/seonho/clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
36
+ "codePath": "clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
37
+ "codePathLocal": "clvla/lerobot_cl/src/lerobot/scripts/train_groot_insight_yaml.py",
38
+ "email": "dankk98@gmail.com",
39
+ "root": "/home/seonho/insight/outputs/dit_only_phase1_ckpt",
40
+ "host": "worker1",
41
+ "executable": "/home/seonho/miniconda3/envs/minje/bin/python3.12",
42
+ "cpu_count": 32,
43
+ "cpu_count_logical": 64,
44
+ "gpu": "NVIDIA A100 80GB PCIe",
45
+ "gpu_count": 4,
46
+ "disk": {
47
+ "/": {
48
+ "total": "3778130485248",
49
+ "used": "3442052386816"
50
+ }
51
+ },
52
+ "memory": {
53
+ "total": "1081620840448"
54
+ },
55
+ "gpu_nvidia": [
56
+ {
57
+ "name": "NVIDIA A100 80GB PCIe",
58
+ "memoryTotal": "85899345920",
59
+ "cudaCores": 6912,
60
+ "architecture": "Ampere",
61
+ "uuid": "GPU-0dd6d491-2dd8-24ac-bb27-872483ab77e9"
62
+ },
63
+ {
64
+ "name": "NVIDIA A100 80GB PCIe",
65
+ "memoryTotal": "85899345920",
66
+ "cudaCores": 6912,
67
+ "architecture": "Ampere",
68
+ "uuid": "GPU-dbfb7925-aec1-50fb-039a-3a4c0a4b2caa"
69
+ },
70
+ {
71
+ "name": "NVIDIA A100 80GB PCIe",
72
+ "memoryTotal": "85899345920",
73
+ "cudaCores": 6912,
74
+ "architecture": "Ampere",
75
+ "uuid": "GPU-ca1c0319-e178-a3c3-ac41-386c2723deac"
76
+ },
77
+ {
78
+ "name": "NVIDIA A100 80GB PCIe",
79
+ "memoryTotal": "85899345920",
80
+ "cudaCores": 6912,
81
+ "architecture": "Ampere",
82
+ "uuid": "GPU-47b6036e-4fde-e399-5515-f73eed752c6d"
83
+ }
84
+ ],
85
+ "cudaVersion": "12.2",
86
+ "writerId": "ddeaytz4ga2t8xmz3osihmmadhyuvdau"
87
+ }
wandb/run-20260518_153314-yn3lgfkj/files/wandb-summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"_runtime":152,"params/whole_total":2427681217,"params/whole_trainable":840593408,"_wandb":{"runtime":152}}
wandb/run-20260518_153314-yn3lgfkj/logs/debug-core.log ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-05-18T15:33:14.368029784+09:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpt29ma1tu/port-3572879.txt","pid":3572879,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
2
+ {"time":"2026-05-18T15:33:14.368891132+09:00","level":"INFO","msg":"server: will exit if parent process dies","ppid":3572879}
3
+ {"time":"2026-05-18T15:33:14.368907866+09:00","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-3572879-3579183-3991850837/socket","Net":"unix"}}
4
+ {"time":"2026-05-18T15:33:14.552175752+09:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
5
+ {"time":"2026-05-18T15:33:14.568692696+09:00","level":"INFO","msg":"handleInformInit: received","streamId":"yn3lgfkj","id":"1(@)"}
6
+ {"time":"2026-05-18T15:33:14.955750856+09:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"yn3lgfkj","id":"1(@)"}
7
+ {"time":"2026-05-18T15:33:20.635218256+09:00","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"oxmrotilaa4a"}
8
+ {"time":"2026-05-18T15:35:48.036183042+09:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
9
+ {"time":"2026-05-18T15:35:48.036225497+09:00","level":"INFO","msg":"connection: closing","id":"1(@)"}
10
+ {"time":"2026-05-18T15:35:48.036239979+09:00","level":"INFO","msg":"server is shutting down"}
11
+ {"time":"2026-05-18T15:35:48.036291142+09:00","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
12
+ {"time":"2026-05-18T15:35:48.036309848+09:00","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-3572879-3579183-3991850837/socket","Net":"unix"}}
wandb/run-20260518_153314-yn3lgfkj/logs/debug-internal.log ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"time":"2026-05-18T15:33:14.568787835+09:00","level":"INFO","msg":"stream: starting","core version":"0.24.2"}
2
+ {"time":"2026-05-18T15:33:14.955449815+09:00","level":"INFO","msg":"stream: created new stream","id":"yn3lgfkj"}
3
+ {"time":"2026-05-18T15:33:14.955563765+09:00","level":"INFO","msg":"handler: started","stream_id":"yn3lgfkj"}
4
+ {"time":"2026-05-18T15:33:14.955729554+09:00","level":"INFO","msg":"stream: started","id":"yn3lgfkj"}
5
+ {"time":"2026-05-18T15:33:14.955755456+09:00","level":"INFO","msg":"writer: started","stream_id":"yn3lgfkj"}
6
+ {"time":"2026-05-18T15:33:14.955762385+09:00","level":"INFO","msg":"sender: started","stream_id":"yn3lgfkj"}
7
+ {"time":"2026-05-18T15:35:48.03624071+09:00","level":"INFO","msg":"stream: closing","id":"yn3lgfkj"}
8
+ {"time":"2026-05-18T15:35:48.678372175+09:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
9
+ {"time":"2026-05-18T15:35:48.977410208+09:00","level":"INFO","msg":"handler: closed","stream_id":"yn3lgfkj"}
10
+ {"time":"2026-05-18T15:35:48.977614839+09:00","level":"INFO","msg":"sender: closed","stream_id":"yn3lgfkj"}
11
+ {"time":"2026-05-18T15:35:48.977668037+09:00","level":"INFO","msg":"stream: closed","id":"yn3lgfkj"}