Hugging Face's logo

chomeed
/
rewind_peg_insertion

Robotics
LeRobot
Safetensors
rewind
Model card Files
xet
 Community
chomeed commited on
Commit
491a37d
·
verified ·
1 Parent(s): 39c943a

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +62 -0
  2. config.json +69 -0
  3. model.safetensors +3 -0
  4. train_config.json +364 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ datasets: chomeed/peg_insertion_success_1000_for_rewind
3
+ library_name: lerobot
4
+ license: apache-2.0
5
+ model_name: rewind
6
+ pipeline_tag: robotics
7
+ tags:
8
+ - robotics
9
+ - lerobot
10
+ - rewind
11
+ ---
12
+
13
+ # Model Card for rewind
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+ _Model type not recognized — please update this template._
19
+
20
+
21
+ This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
22
+ See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
23
+
24
+ ---
25
+
26
+ ## How to Get Started with the Model
27
+
28
+ For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
29
+ Below is the short version on how to train and run inference/eval:
30
+
31
+ ### Train from scratch
32
+
33
+ ```bash
34
+ lerobot-train \
35
+ --dataset.repo_id=${HF_USER}/<dataset> \
36
+ --policy.type=act \
37
+ --output_dir=outputs/train/<desired_policy_repo_id> \
38
+ --job_name=lerobot_training \
39
+ --policy.device=cuda \
40
+ --policy.repo_id=${HF_USER}/<desired_policy_repo_id>
41
+ --wandb.enable=true
42
+ ```
43
+
44
+ _Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
45
+
46
+ ### Evaluate the policy/run inference
47
+
48
+ ```bash
49
+ lerobot-record \
50
+ --robot.type=so100_follower \
51
+ --dataset.repo_id=<hf_user>/eval_<dataset> \
52
+ --policy.path=<hf_user>/<desired_policy_repo_id> \
53
+ --episodes=10
54
+ ```
55
+
56
+ Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
57
+
58
+ ---
59
+
60
+ ## Model Details
61
+
62
+ - **License:** apache-2.0
config.json ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "rewind",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.front_rgb": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 256
11
+ ]
12
+ },
13
+ "observation.images.wrist_rgb": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 256
19
+ ]
20
+ },
21
+ "observation.dinov2.front_rgb": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 768
25
+ ]
26
+ },
27
+ "observation.dinov2.wrist_rgb": {
28
+ "type": "STATE",
29
+ "shape": [
30
+ 768
31
+ ]
32
+ }
33
+ },
34
+ "output_features": {},
35
+ "device": "cuda",
36
+ "use_amp": false,
37
+ "use_peft": false,
38
+ "push_to_hub": true,
39
+ "repo_id": "chomeed/rewind_peg_insertion",
40
+ "private": null,
41
+ "tags": null,
42
+ "license": null,
43
+ "pretrained_path": null,
44
+ "d_model": 512,
45
+ "vis_emb_dim": 768,
46
+ "text_emb_dim": 384,
47
+ "n_layers": 4,
48
+ "n_heads": 8,
49
+ "dropout": 0.1,
50
+ "num_cameras": 2,
51
+ "optimizer_lr": 0.0001,
52
+ "optimizer_betas": [
53
+ 0.9,
54
+ 0.95
55
+ ],
56
+ "optimizer_eps": 1e-08,
57
+ "optimizer_weight_decay": 1e-10,
58
+ "optimizer_grad_clip_norm": 10.0,
59
+ "scheduler_warmup_steps": 1000,
60
+ "scheduler_decay_steps": 5000,
61
+ "scheduler_decay_lr": 2.5e-06,
62
+ "scheduler_peak_lr": 5e-05,
63
+ "normalization_mapping": {
64
+ "VISUAL": "IDENTITY",
65
+ "STATE": "MEAN_STD",
66
+ "LANGUAGE": "IDENTITY",
67
+ "REWARD": "IDENTITY"
68
+ }
69
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc84040e1935f3a6ee3f328cffd1ab691b65bd0c5e1add36cc6c37f10f0f0c82
3
+ size 55971724
train_config.json ADDED
@@ -0,0 +1,364 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "chomeed/peg_insertion_success_1000_for_rewind",
4
+ "root": null,
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",
86
+ "n_obs_steps": 1,
87
+ "input_features": {
88
+ "observation.images.front_rgb": {
89
+ "type": "VISUAL",
90
+ "shape": [
91
+ 3,
92
+ 256,
93
+ 256
94
+ ]
95
+ },
96
+ "observation.images.wrist_rgb": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 3,
100
+ 256,
101
+ 256
102
+ ]
103
+ },
104
+ "observation.dinov2.front_rgb": {
105
+ "type": "STATE",
106
+ "shape": [
107
+ 768
108
+ ]
109
+ },
110
+ "observation.dinov2.wrist_rgb": {
111
+ "type": "STATE",
112
+ "shape": [
113
+ 768
114
+ ]
115
+ }
116
+ },
117
+ "output_features": {},
118
+ "device": "cuda",
119
+ "use_amp": false,
120
+ "use_peft": false,
121
+ "push_to_hub": true,
122
+ "repo_id": "chomeed/rewind_peg_insertion",
123
+ "private": null,
124
+ "tags": null,
125
+ "license": null,
126
+ "pretrained_path": null,
127
+ "d_model": 512,
128
+ "vis_emb_dim": 768,
129
+ "text_emb_dim": 384,
130
+ "n_layers": 4,
131
+ "n_heads": 8,
132
+ "dropout": 0.1,
133
+ "num_cameras": 2,
134
+ "optimizer_lr": 0.0001,
135
+ "optimizer_betas": [
136
+ 0.9,
137
+ 0.95
138
+ ],
139
+ "optimizer_eps": 1e-08,
140
+ "optimizer_weight_decay": 1e-10,
141
+ "optimizer_grad_clip_norm": 10.0,
142
+ "scheduler_warmup_steps": 1000,
143
+ "scheduler_decay_steps": 5000,
144
+ "scheduler_decay_lr": 2.5e-06,
145
+ "scheduler_peak_lr": 5e-05,
146
+ "normalization_mapping": {
147
+ "VISUAL": "IDENTITY",
148
+ "STATE": "MEAN_STD",
149
+ "LANGUAGE": "IDENTITY",
150
+ "REWARD": "IDENTITY"
151
+ }
152
+ },
153
+ "output_dir": "outputs/train/2026-04-14/14-29-48_rewind_peg_insertion",
154
+ "job_name": "rewind_peg_insertion",
155
+ "resume": false,
156
+ "seed": 1000,
157
+ "num_workers": 16,
158
+ "batch_size": 16,
159
+ "steps": 5000,
160
+ "eval_freq": 500,
161
+ "log_freq": 50,
162
+ "tolerance_s": 0.0001,
163
+ "save_checkpoint": true,
164
+ "save_freq": 500,
165
+ "use_policy_training_preset": true,
166
+ "optimizer": {
167
+ "type": "adamw",
168
+ "lr": 0.0001,
169
+ "weight_decay": 1e-10,
170
+ "grad_clip_norm": 10.0,
171
+ "betas": [
172
+ 0.9,
173
+ 0.95
174
+ ],
175
+ "eps": 1e-08
176
+ },
177
+ "scheduler": {
178
+ "type": "cosine_decay_with_warmup",
179
+ "num_warmup_steps": 1000,
180
+ "num_decay_steps": 5000,
181
+ "peak_lr": 5e-05,
182
+ "decay_lr": 2.5e-06
183
+ },
184
+ "val": {
185
+ "repo_id": "chomeed/peg_insertion_success_1000_for_rewind",
186
+ "root": null,
187
+ "episodes": null,
188
+ "image_transforms": {
189
+ "enable": false,
190
+ "max_num_transforms": 3,
191
+ "random_order": false,
192
+ "tfs": {
193
+ "brightness": {
194
+ "weight": 1.0,
195
+ "type": "ColorJitter",
196
+ "kwargs": {
197
+ "brightness": [
198
+ 0.8,
199
+ 1.2
200
+ ]
201
+ }
202
+ },
203
+ "contrast": {
204
+ "weight": 1.0,
205
+ "type": "ColorJitter",
206
+ "kwargs": {
207
+ "contrast": [
208
+ 0.8,
209
+ 1.2
210
+ ]
211
+ }
212
+ },
213
+ "saturation": {
214
+ "weight": 1.0,
215
+ "type": "ColorJitter",
216
+ "kwargs": {
217
+ "saturation": [
218
+ 0.5,
219
+ 1.5
220
+ ]
221
+ }
222
+ },
223
+ "hue": {
224
+ "weight": 1.0,
225
+ "type": "ColorJitter",
226
+ "kwargs": {
227
+ "hue": [
228
+ -0.05,
229
+ 0.05
230
+ ]
231
+ }
232
+ },
233
+ "sharpness": {
234
+ "weight": 1.0,
235
+ "type": "SharpnessJitter",
236
+ "kwargs": {
237
+ "sharpness": [
238
+ 0.5,
239
+ 1.5
240
+ ]
241
+ }
242
+ },
243
+ "affine": {
244
+ "weight": 1.0,
245
+ "type": "RandomAffine",
246
+ "kwargs": {
247
+ "degrees": [
248
+ -5.0,
249
+ 5.0
250
+ ],
251
+ "translate": [
252
+ 0.05,
253
+ 0.05
254
+ ]
255
+ }
256
+ }
257
+ }
258
+ },
259
+ "revision": null,
260
+ "use_imagenet_stats": true,
261
+ "video_backend": "torchcodec",
262
+ "streaming": false
263
+ },
264
+ "test": {
265
+ "repo_id": [
266
+ "chomeed/aloha_static_coffee_single",
267
+ "jellyho/aloha_handover_box_single"
268
+ ],
269
+ "root": null,
270
+ "episodes": null,
271
+ "image_transforms": {
272
+ "enable": false,
273
+ "max_num_transforms": 3,
274
+ "random_order": false,
275
+ "tfs": {
276
+ "brightness": {
277
+ "weight": 1.0,
278
+ "type": "ColorJitter",
279
+ "kwargs": {
280
+ "brightness": [
281
+ 0.8,
282
+ 1.2
283
+ ]
284
+ }
285
+ },
286
+ "contrast": {
287
+ "weight": 1.0,
288
+ "type": "ColorJitter",
289
+ "kwargs": {
290
+ "contrast": [
291
+ 0.8,
292
+ 1.2
293
+ ]
294
+ }
295
+ },
296
+ "saturation": {
297
+ "weight": 1.0,
298
+ "type": "ColorJitter",
299
+ "kwargs": {
300
+ "saturation": [
301
+ 0.5,
302
+ 1.5
303
+ ]
304
+ }
305
+ },
306
+ "hue": {
307
+ "weight": 1.0,
308
+ "type": "ColorJitter",
309
+ "kwargs": {
310
+ "hue": [
311
+ -0.05,
312
+ 0.05
313
+ ]
314
+ }
315
+ },
316
+ "sharpness": {
317
+ "weight": 1.0,
318
+ "type": "SharpnessJitter",
319
+ "kwargs": {
320
+ "sharpness": [
321
+ 0.5,
322
+ 1.5
323
+ ]
324
+ }
325
+ },
326
+ "affine": {
327
+ "weight": 1.0,
328
+ "type": "RandomAffine",
329
+ "kwargs": {
330
+ "degrees": [
331
+ -5.0,
332
+ 5.0
333
+ ],
334
+ "translate": [
335
+ 0.05,
336
+ 0.05
337
+ ]
338
+ }
339
+ }
340
+ }
341
+ },
342
+ "revision": null,
343
+ "use_imagenet_stats": true,
344
+ "video_backend": "torchcodec",
345
+ "streaming": false
346
+ },
347
+ "wandb": {
348
+ "enable": false,
349
+ "disable_artifact": true,
350
+ "project": "long_horizon_reward_modeling",
351
+ "entity": null,
352
+ "notes": null,
353
+ "run_id": null,
354
+ "mode": null
355
+ },
356
+ "peft": null,
357
+ "use_rabc": false,
358
+ "rabc_progress_path": null,
359
+ "rabc_kappa": 0.01,
360
+ "rabc_epsilon": 1e-06,
361
+ "rabc_head_mode": "sparse",
362
+ "rename_map": {},
363
+ "checkpoint_path": null
364
+ }