Robotics
LeRobot
Safetensors
act
jayyucippg commited on
Commit
17e0ef8
·
verified ·
1 Parent(s): 1ad8bc0

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +62 -0
  2. config.json +61 -0
  3. model.safetensors +3 -0
  4. train_config.json +376 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ datasets: jayyucippg/merge-aug
3
+ library_name: lerobot
4
+ license: apache-2.0
5
+ model_name: act
6
+ pipeline_tag: robotics
7
+ tags:
8
+ - act
9
+ - robotics
10
+ - lerobot
11
+ ---
12
+
13
+ # Model Card for act
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+ [Action Chunking with Transformers (ACT)](https://huggingface.co/papers/2304.13705) is an imitation-learning method that predicts short action chunks instead of single steps. It learns from teleoperated data and often achieves high success rates.
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
+ python -m lerobot.scripts.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
+ python -m 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 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "normalization_mapping": {
5
+ "VISUAL": "MEAN_STD",
6
+ "STATE": "MEAN_STD",
7
+ "ACTION": "MEAN_STD"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 6
14
+ ]
15
+ },
16
+ "observation.images.side": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ }
24
+ },
25
+ "output_features": {
26
+ "action": {
27
+ "type": "ACTION",
28
+ "shape": [
29
+ 6
30
+ ]
31
+ }
32
+ },
33
+ "device": "cuda",
34
+ "use_amp": false,
35
+ "push_to_hub": true,
36
+ "repo_id": "jayyucippg/act-merge-aug",
37
+ "private": null,
38
+ "tags": null,
39
+ "license": null,
40
+ "chunk_size": 100,
41
+ "n_action_steps": 100,
42
+ "vision_backbone": "resnet18",
43
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
44
+ "replace_final_stride_with_dilation": false,
45
+ "pre_norm": false,
46
+ "dim_model": 512,
47
+ "n_heads": 8,
48
+ "dim_feedforward": 3200,
49
+ "feedforward_activation": "relu",
50
+ "n_encoder_layers": 4,
51
+ "n_decoder_layers": 1,
52
+ "use_vae": true,
53
+ "latent_dim": 32,
54
+ "n_vae_encoder_layers": 4,
55
+ "temporal_ensemble_coeff": null,
56
+ "dropout": 0.1,
57
+ "kl_weight": 10.0,
58
+ "optimizer_lr": 1e-05,
59
+ "optimizer_weight_decay": 0.0001,
60
+ "optimizer_lr_backbone": 1e-05
61
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7715d53530992663ef81692586c3b983f970d3bc7181f10f00345666621f78e
3
+ size 206700792
train_config.json ADDED
@@ -0,0 +1,376 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "jayyucippg/merge-aug",
4
+ "root": "/mnt/hdd/huggingface/lerobot/jayyucippg/merge-aug",
5
+ "episodes": [
6
+ 0,
7
+ 1,
8
+ 2,
9
+ 3,
10
+ 4,
11
+ 5,
12
+ 6,
13
+ 7,
14
+ 8,
15
+ 9,
16
+ 10,
17
+ 11,
18
+ 12,
19
+ 13,
20
+ 14,
21
+ 15,
22
+ 16,
23
+ 17,
24
+ 18,
25
+ 19,
26
+ 20,
27
+ 21,
28
+ 22,
29
+ 23,
30
+ 24,
31
+ 25,
32
+ 26,
33
+ 27,
34
+ 28,
35
+ 29,
36
+ 30,
37
+ 31,
38
+ 32,
39
+ 33,
40
+ 34,
41
+ 35,
42
+ 36,
43
+ 37,
44
+ 38,
45
+ 39,
46
+ 40,
47
+ 42,
48
+ 43,
49
+ 44,
50
+ 45,
51
+ 46,
52
+ 47,
53
+ 48,
54
+ 49,
55
+ 50,
56
+ 51,
57
+ 52,
58
+ 53,
59
+ 54,
60
+ 55,
61
+ 56,
62
+ 57,
63
+ 58,
64
+ 59,
65
+ 60,
66
+ 61,
67
+ 62,
68
+ 63,
69
+ 64,
70
+ 65,
71
+ 66,
72
+ 67,
73
+ 68,
74
+ 69,
75
+ 70,
76
+ 71,
77
+ 72,
78
+ 73,
79
+ 74,
80
+ 75,
81
+ 76,
82
+ 78,
83
+ 79,
84
+ 80,
85
+ 81,
86
+ 82,
87
+ 83,
88
+ 84,
89
+ 85,
90
+ 86,
91
+ 87,
92
+ 88,
93
+ 89,
94
+ 90,
95
+ 91,
96
+ 92,
97
+ 93,
98
+ 94,
99
+ 95,
100
+ 96,
101
+ 97,
102
+ 98,
103
+ 99,
104
+ 100,
105
+ 101,
106
+ 102,
107
+ 103,
108
+ 104,
109
+ 105,
110
+ 106,
111
+ 107,
112
+ 108,
113
+ 109,
114
+ 110,
115
+ 111,
116
+ 112,
117
+ 113,
118
+ 114,
119
+ 115,
120
+ 116,
121
+ 117,
122
+ 118,
123
+ 119,
124
+ 120,
125
+ 121,
126
+ 122,
127
+ 123,
128
+ 124,
129
+ 125,
130
+ 126,
131
+ 127,
132
+ 128,
133
+ 129,
134
+ 130,
135
+ 131,
136
+ 132,
137
+ 133,
138
+ 134,
139
+ 135,
140
+ 136,
141
+ 137,
142
+ 138,
143
+ 139,
144
+ 140,
145
+ 141,
146
+ 142,
147
+ 143,
148
+ 144,
149
+ 145,
150
+ 146,
151
+ 147,
152
+ 148,
153
+ 149,
154
+ 150,
155
+ 151,
156
+ 152,
157
+ 153,
158
+ 154,
159
+ 155,
160
+ 156,
161
+ 157,
162
+ 158,
163
+ 159,
164
+ 160,
165
+ 161,
166
+ 162,
167
+ 163,
168
+ 164,
169
+ 165,
170
+ 166,
171
+ 167,
172
+ 168,
173
+ 169,
174
+ 170,
175
+ 171,
176
+ 172,
177
+ 173,
178
+ 174,
179
+ 175,
180
+ 176,
181
+ 177,
182
+ 178,
183
+ 179,
184
+ 180,
185
+ 181,
186
+ 182,
187
+ 183,
188
+ 184,
189
+ 185,
190
+ 186,
191
+ 187,
192
+ 188,
193
+ 189,
194
+ 190,
195
+ 191,
196
+ 192,
197
+ 193,
198
+ 194,
199
+ 195,
200
+ 196,
201
+ 197,
202
+ 198,
203
+ 199,
204
+ 200,
205
+ 201,
206
+ 202,
207
+ 203,
208
+ 204,
209
+ 205,
210
+ 206,
211
+ 207,
212
+ 208,
213
+ 209
214
+ ],
215
+ "image_transforms": {
216
+ "enable": false,
217
+ "max_num_transforms": 3,
218
+ "random_order": false,
219
+ "tfs": {
220
+ "brightness": {
221
+ "weight": 1.0,
222
+ "type": "ColorJitter",
223
+ "kwargs": {
224
+ "brightness": [
225
+ 0.8,
226
+ 1.2
227
+ ]
228
+ }
229
+ },
230
+ "contrast": {
231
+ "weight": 1.0,
232
+ "type": "ColorJitter",
233
+ "kwargs": {
234
+ "contrast": [
235
+ 0.8,
236
+ 1.2
237
+ ]
238
+ }
239
+ },
240
+ "saturation": {
241
+ "weight": 1.0,
242
+ "type": "ColorJitter",
243
+ "kwargs": {
244
+ "saturation": [
245
+ 0.5,
246
+ 1.5
247
+ ]
248
+ }
249
+ },
250
+ "hue": {
251
+ "weight": 1.0,
252
+ "type": "ColorJitter",
253
+ "kwargs": {
254
+ "hue": [
255
+ -0.05,
256
+ 0.05
257
+ ]
258
+ }
259
+ },
260
+ "sharpness": {
261
+ "weight": 1.0,
262
+ "type": "SharpnessJitter",
263
+ "kwargs": {
264
+ "sharpness": [
265
+ 0.5,
266
+ 1.5
267
+ ]
268
+ }
269
+ }
270
+ }
271
+ },
272
+ "revision": null,
273
+ "use_imagenet_stats": true,
274
+ "video_backend": "torchcodec"
275
+ },
276
+ "env": null,
277
+ "policy": {
278
+ "type": "act",
279
+ "n_obs_steps": 1,
280
+ "normalization_mapping": {
281
+ "VISUAL": "MEAN_STD",
282
+ "STATE": "MEAN_STD",
283
+ "ACTION": "MEAN_STD"
284
+ },
285
+ "input_features": {
286
+ "observation.state": {
287
+ "type": "STATE",
288
+ "shape": [
289
+ 6
290
+ ]
291
+ },
292
+ "observation.images.side": {
293
+ "type": "VISUAL",
294
+ "shape": [
295
+ 3,
296
+ 480,
297
+ 640
298
+ ]
299
+ }
300
+ },
301
+ "output_features": {
302
+ "action": {
303
+ "type": "ACTION",
304
+ "shape": [
305
+ 6
306
+ ]
307
+ }
308
+ },
309
+ "device": "cuda",
310
+ "use_amp": false,
311
+ "push_to_hub": true,
312
+ "repo_id": "jayyucippg/act-merge-aug",
313
+ "private": null,
314
+ "tags": null,
315
+ "license": null,
316
+ "chunk_size": 100,
317
+ "n_action_steps": 100,
318
+ "vision_backbone": "resnet18",
319
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
320
+ "replace_final_stride_with_dilation": false,
321
+ "pre_norm": false,
322
+ "dim_model": 512,
323
+ "n_heads": 8,
324
+ "dim_feedforward": 3200,
325
+ "feedforward_activation": "relu",
326
+ "n_encoder_layers": 4,
327
+ "n_decoder_layers": 1,
328
+ "use_vae": true,
329
+ "latent_dim": 32,
330
+ "n_vae_encoder_layers": 4,
331
+ "temporal_ensemble_coeff": null,
332
+ "dropout": 0.1,
333
+ "kl_weight": 10.0,
334
+ "optimizer_lr": 1e-05,
335
+ "optimizer_weight_decay": 0.0001,
336
+ "optimizer_lr_backbone": 1e-05
337
+ },
338
+ "output_dir": "/mnt/hdd/jyq/tmp/train/merge-aug-1",
339
+ "job_name": "act-merge-aug",
340
+ "resume": false,
341
+ "seed": 1000,
342
+ "num_workers": 4,
343
+ "batch_size": 8,
344
+ "steps": 100000,
345
+ "eval_freq": 20000,
346
+ "log_freq": 200,
347
+ "save_checkpoint": true,
348
+ "save_freq": 20000,
349
+ "use_policy_training_preset": true,
350
+ "optimizer": {
351
+ "type": "adamw",
352
+ "lr": 1e-05,
353
+ "weight_decay": 0.0001,
354
+ "grad_clip_norm": 10.0,
355
+ "betas": [
356
+ 0.9,
357
+ 0.999
358
+ ],
359
+ "eps": 1e-08
360
+ },
361
+ "scheduler": null,
362
+ "eval": {
363
+ "n_episodes": 50,
364
+ "batch_size": 50,
365
+ "use_async_envs": false
366
+ },
367
+ "wandb": {
368
+ "enable": true,
369
+ "disable_artifact": false,
370
+ "project": "lerobot",
371
+ "entity": null,
372
+ "notes": null,
373
+ "run_id": "sstjo4pm",
374
+ "mode": null
375
+ }
376
+ }