Hugging Face's logo

chomeed
/
rewind_state_dexmg_two_arm_threading

Robotics
LeRobot
Safetensors
reward-model
rewind
progress-reward
dexmimicgen
bimanual
Model card Files
xet
 Community
chomeed commited on
Commit
9d73fa8
·
verified ·
1 Parent(s): e4fcc88

Upload checkpoint from /lustre/chomeed/exaflops/data/ckpts/rewind/rewind_state_dexmg_threading_v1/checkpoints/last/pretrained_model

Browse files
Files changed (7) hide show
  1. config.json +88 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +33 -0
  4. policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. policy_preprocessor.json +57 -0
  6. policy_preprocessor_step_5_normalizer_processor.safetensors +3 -0
  7. train_config.json +383 -0
config.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "rewind_state",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.agentview": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 84,
10
+ 84
11
+ ]
12
+ },
13
+ "observation.images.robot0_eye_in_hand": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 84,
18
+ 84
19
+ ]
20
+ },
21
+ "observation.images.robot1_eye_in_hand": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 84,
26
+ 84
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 18
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 14
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": false,
46
+ "use_peft": false,
47
+ "push_to_hub": false,
48
+ "repo_id": "chomeed/rewind_state_dexmg_threading_v1",
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": null,
53
+ "d_model": 512,
54
+ "vis_emb_dim": 768,
55
+ "text_emb_dim": 384,
56
+ "state_dim": 18,
57
+ "n_layers": 4,
58
+ "n_heads": 8,
59
+ "dropout": 0.1,
60
+ "num_cameras": 3,
61
+ "finetune_encoder": true,
62
+ "finetune_encoder_blocks": 2,
63
+ "encoder_ckpt": "facebook/dinov2-base",
64
+ "encoder_lr": 1e-05,
65
+ "encoder_camera_keys": [
66
+ "observation.images.agentview",
67
+ "observation.images.robot0_eye_in_hand",
68
+ "observation.images.robot1_eye_in_hand"
69
+ ],
70
+ "optimizer_lr": 0.0001,
71
+ "optimizer_betas": [
72
+ 0.9,
73
+ 0.95
74
+ ],
75
+ "optimizer_eps": 1e-08,
76
+ "optimizer_weight_decay": 1e-10,
77
+ "optimizer_grad_clip_norm": 10.0,
78
+ "scheduler_warmup_steps": 1000,
79
+ "scheduler_decay_steps": 50000,
80
+ "scheduler_decay_lr": 2.5e-06,
81
+ "scheduler_peak_lr": 5e-05,
82
+ "normalization_mapping": {
83
+ "VISUAL": "IDENTITY",
84
+ "STATE": "MEAN_STD",
85
+ "LANGUAGE": "IDENTITY",
86
+ "REWARD": "IDENTITY"
87
+ }
88
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8dad424e31cfb3e6a0b896cf1e00459137684a906505fb781019f390f47a3bdc
3
+ size 411817036
policy_postprocessor.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "unnormalizer_processor",
6
+ "config": {
7
+ "eps": 1e-08,
8
+ "features": {
9
+ "action": {
10
+ "type": "ACTION",
11
+ "shape": [
12
+ 14
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "IDENTITY",
18
+ "STATE": "MEAN_STD",
19
+ "LANGUAGE": "IDENTITY",
20
+ "REWARD": "IDENTITY"
21
+ }
22
+ },
23
+ "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
24
+ },
25
+ {
26
+ "registry_name": "device_processor",
27
+ "config": {
28
+ "device": "cpu",
29
+ "float_dtype": null
30
+ }
31
+ }
32
+ ]
33
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c071b13a191d6a68af4ef1037cd5548b0f9a04fabeeb71f9cf323788851b290
3
+ size 11056
policy_preprocessor.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "class": "lerobot_policy_rewind_state.processor_rewind_state.RawImageLengthsProcessorStep",
16
+ "config": {}
17
+ },
18
+ {
19
+ "class": "lerobot_policy_rewind_state.processor_rewind_state.MiniLMTextEmbeddingProcessorStep",
20
+ "config": {}
21
+ },
22
+ {
23
+ "registry_name": "device_processor",
24
+ "config": {
25
+ "device": "cuda",
26
+ "float_dtype": null
27
+ }
28
+ },
29
+ {
30
+ "registry_name": "normalizer_processor",
31
+ "config": {
32
+ "eps": 1e-08,
33
+ "features": {
34
+ "observation.state": {
35
+ "type": "STATE",
36
+ "shape": [
37
+ 18
38
+ ]
39
+ },
40
+ "action": {
41
+ "type": "ACTION",
42
+ "shape": [
43
+ 14
44
+ ]
45
+ }
46
+ },
47
+ "norm_map": {
48
+ "VISUAL": "IDENTITY",
49
+ "STATE": "MEAN_STD",
50
+ "LANGUAGE": "IDENTITY",
51
+ "REWARD": "IDENTITY"
52
+ }
53
+ },
54
+ "state_file": "policy_preprocessor_step_5_normalizer_processor.safetensors"
55
+ }
56
+ ]
57
+ }
policy_preprocessor_step_5_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c071b13a191d6a68af4ef1037cd5548b0f9a04fabeeb71f9cf323788851b290
3
+ size 11056
train_config.json ADDED
@@ -0,0 +1,383 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "chomeed/dexmg_two_arm_threading",
4
+ "root": "/home/chomeed/workspace/proj-rl-finetuning/lerobot_dataset/chomeed/dexmg_two_arm_threading",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "streaming": false,
81
+ "cf_aug": false
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "rewind_state",
86
+ "n_obs_steps": 1,
87
+ "input_features": {
88
+ "observation.images.agentview": {
89
+ "type": "VISUAL",
90
+ "shape": [
91
+ 3,
92
+ 84,
93
+ 84
94
+ ]
95
+ },
96
+ "observation.images.robot0_eye_in_hand": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 3,
100
+ 84,
101
+ 84
102
+ ]
103
+ },
104
+ "observation.images.robot1_eye_in_hand": {
105
+ "type": "VISUAL",
106
+ "shape": [
107
+ 3,
108
+ 84,
109
+ 84
110
+ ]
111
+ },
112
+ "observation.state": {
113
+ "type": "STATE",
114
+ "shape": [
115
+ 18
116
+ ]
117
+ }
118
+ },
119
+ "output_features": {
120
+ "action": {
121
+ "type": "ACTION",
122
+ "shape": [
123
+ 14
124
+ ]
125
+ }
126
+ },
127
+ "device": "cuda",
128
+ "use_amp": false,
129
+ "use_peft": false,
130
+ "push_to_hub": false,
131
+ "repo_id": "chomeed/rewind_state_dexmg_threading_v1",
132
+ "private": null,
133
+ "tags": null,
134
+ "license": null,
135
+ "pretrained_path": null,
136
+ "d_model": 512,
137
+ "vis_emb_dim": 768,
138
+ "text_emb_dim": 384,
139
+ "state_dim": 18,
140
+ "n_layers": 4,
141
+ "n_heads": 8,
142
+ "dropout": 0.1,
143
+ "num_cameras": 3,
144
+ "finetune_encoder": true,
145
+ "finetune_encoder_blocks": 2,
146
+ "encoder_ckpt": "facebook/dinov2-base",
147
+ "encoder_lr": 1e-05,
148
+ "encoder_camera_keys": [
149
+ "observation.images.agentview",
150
+ "observation.images.robot0_eye_in_hand",
151
+ "observation.images.robot1_eye_in_hand"
152
+ ],
153
+ "optimizer_lr": 0.0001,
154
+ "optimizer_betas": [
155
+ 0.9,
156
+ 0.95
157
+ ],
158
+ "optimizer_eps": 1e-08,
159
+ "optimizer_weight_decay": 1e-10,
160
+ "optimizer_grad_clip_norm": 10.0,
161
+ "scheduler_warmup_steps": 1000,
162
+ "scheduler_decay_steps": 50000,
163
+ "scheduler_decay_lr": 2.5e-06,
164
+ "scheduler_peak_lr": 5e-05,
165
+ "normalization_mapping": {
166
+ "VISUAL": "IDENTITY",
167
+ "STATE": "MEAN_STD",
168
+ "LANGUAGE": "IDENTITY",
169
+ "REWARD": "IDENTITY"
170
+ }
171
+ },
172
+ "output_dir": "/scratch2/chomeed/outputs/rewind/rewind_state_dexmg_threading_v1",
173
+ "job_name": "rewind_state_dexmg_threading_v1",
174
+ "resume": false,
175
+ "seed": 1000,
176
+ "num_workers": 8,
177
+ "batch_size": 16,
178
+ "steps": 50000,
179
+ "eval_freq": 1000,
180
+ "log_freq": 100,
181
+ "tolerance_s": 0.0001,
182
+ "save_checkpoint": true,
183
+ "save_freq": 1000,
184
+ "use_policy_training_preset": true,
185
+ "optimizer": {
186
+ "type": "adamw",
187
+ "lr": 0.0001,
188
+ "weight_decay": 1e-10,
189
+ "grad_clip_norm": 10.0,
190
+ "betas": [
191
+ 0.9,
192
+ 0.95
193
+ ],
194
+ "eps": 1e-08
195
+ },
196
+ "scheduler": {
197
+ "type": "cosine_decay_with_warmup",
198
+ "num_warmup_steps": 1000,
199
+ "num_decay_steps": 50000,
200
+ "peak_lr": 5e-05,
201
+ "decay_lr": 2.5e-06
202
+ },
203
+ "val": {
204
+ "repo_id": "chomeed/dexmg_two_arm_threading",
205
+ "root": "/home/chomeed/workspace/proj-rl-finetuning/lerobot_dataset/chomeed/dexmg_two_arm_threading",
206
+ "episodes": null,
207
+ "image_transforms": {
208
+ "enable": false,
209
+ "max_num_transforms": 3,
210
+ "random_order": false,
211
+ "tfs": {
212
+ "brightness": {
213
+ "weight": 1.0,
214
+ "type": "ColorJitter",
215
+ "kwargs": {
216
+ "brightness": [
217
+ 0.8,
218
+ 1.2
219
+ ]
220
+ }
221
+ },
222
+ "contrast": {
223
+ "weight": 1.0,
224
+ "type": "ColorJitter",
225
+ "kwargs": {
226
+ "contrast": [
227
+ 0.8,
228
+ 1.2
229
+ ]
230
+ }
231
+ },
232
+ "saturation": {
233
+ "weight": 1.0,
234
+ "type": "ColorJitter",
235
+ "kwargs": {
236
+ "saturation": [
237
+ 0.5,
238
+ 1.5
239
+ ]
240
+ }
241
+ },
242
+ "hue": {
243
+ "weight": 1.0,
244
+ "type": "ColorJitter",
245
+ "kwargs": {
246
+ "hue": [
247
+ -0.05,
248
+ 0.05
249
+ ]
250
+ }
251
+ },
252
+ "sharpness": {
253
+ "weight": 1.0,
254
+ "type": "SharpnessJitter",
255
+ "kwargs": {
256
+ "sharpness": [
257
+ 0.5,
258
+ 1.5
259
+ ]
260
+ }
261
+ },
262
+ "affine": {
263
+ "weight": 1.0,
264
+ "type": "RandomAffine",
265
+ "kwargs": {
266
+ "degrees": [
267
+ -5.0,
268
+ 5.0
269
+ ],
270
+ "translate": [
271
+ 0.05,
272
+ 0.05
273
+ ]
274
+ }
275
+ }
276
+ }
277
+ },
278
+ "revision": null,
279
+ "use_imagenet_stats": true,
280
+ "video_backend": "torchcodec",
281
+ "streaming": false
282
+ },
283
+ "test": {
284
+ "repo_id": [
285
+ "chomeed/aloha_static_coffee_single",
286
+ "jellyho/aloha_handover_box_single"
287
+ ],
288
+ "root": null,
289
+ "episodes": null,
290
+ "image_transforms": {
291
+ "enable": false,
292
+ "max_num_transforms": 3,
293
+ "random_order": false,
294
+ "tfs": {
295
+ "brightness": {
296
+ "weight": 1.0,
297
+ "type": "ColorJitter",
298
+ "kwargs": {
299
+ "brightness": [
300
+ 0.8,
301
+ 1.2
302
+ ]
303
+ }
304
+ },
305
+ "contrast": {
306
+ "weight": 1.0,
307
+ "type": "ColorJitter",
308
+ "kwargs": {
309
+ "contrast": [
310
+ 0.8,
311
+ 1.2
312
+ ]
313
+ }
314
+ },
315
+ "saturation": {
316
+ "weight": 1.0,
317
+ "type": "ColorJitter",
318
+ "kwargs": {
319
+ "saturation": [
320
+ 0.5,
321
+ 1.5
322
+ ]
323
+ }
324
+ },
325
+ "hue": {
326
+ "weight": 1.0,
327
+ "type": "ColorJitter",
328
+ "kwargs": {
329
+ "hue": [
330
+ -0.05,
331
+ 0.05
332
+ ]
333
+ }
334
+ },
335
+ "sharpness": {
336
+ "weight": 1.0,
337
+ "type": "SharpnessJitter",
338
+ "kwargs": {
339
+ "sharpness": [
340
+ 0.5,
341
+ 1.5
342
+ ]
343
+ }
344
+ },
345
+ "affine": {
346
+ "weight": 1.0,
347
+ "type": "RandomAffine",
348
+ "kwargs": {
349
+ "degrees": [
350
+ -5.0,
351
+ 5.0
352
+ ],
353
+ "translate": [
354
+ 0.05,
355
+ 0.05
356
+ ]
357
+ }
358
+ }
359
+ }
360
+ },
361
+ "revision": null,
362
+ "use_imagenet_stats": true,
363
+ "video_backend": "torchcodec",
364
+ "streaming": false
365
+ },
366
+ "wandb": {
367
+ "enable": true,
368
+ "disable_artifact": true,
369
+ "project": "long_horizon_reward_modeling",
370
+ "entity": null,
371
+ "notes": null,
372
+ "run_id": "jkk0epde",
373
+ "mode": null
374
+ },
375
+ "peft": null,
376
+ "use_rabc": false,
377
+ "rabc_progress_path": null,
378
+ "rabc_kappa": 0.01,
379
+ "rabc_epsilon": 1e-06,
380
+ "rabc_head_mode": "sparse",
381
+ "rename_map": {},
382
+ "checkpoint_path": null
383
+ }