Hugging Face's logo

XYZPIT
/
multitask_dit_260518_r1lite_diff

Robotics
LeRobot
Safetensors
multi_task_dit
Model card Files
xet
 Community
2sin0 commited on
Commit
dded95d
·
verified ·
1 Parent(s): a57ed3e

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +62 -0
  2. config.json +235 -0
  3. model.safetensors +3 -0
  4. train_config.json +374 -0
README.md ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ datasets: XYZPIT/260413_r1lite_lerobot
3
+ library_name: lerobot
4
+ license: apache-2.0
5
+ model_name: multi_task_dit
6
+ pipeline_tag: robotics
7
+ tags:
8
+ - multi_task_dit
9
+ - robotics
10
+ - lerobot
11
+ ---
12
+
13
+ # Model Card for multi_task_dit
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 +1,235 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.head_rgb": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 720,
10
+ 1280
11
+ ]
12
+ },
13
+ "observation.images.head_right_rgb": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 720,
18
+ 1280
19
+ ]
20
+ },
21
+ "observation.images.left_wrist_rgb": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 360,
26
+ 640
27
+ ]
28
+ },
29
+ "observation.images.right_wrist_rgb": {
30
+ "type": "VISUAL",
31
+ "shape": [
32
+ 3,
33
+ 360,
34
+ 640
35
+ ]
36
+ },
37
+ "observation.state.left_arm": {
38
+ "type": "STATE",
39
+ "shape": [
40
+ 6
41
+ ]
42
+ },
43
+ "observation.state.left_arm.velocities": {
44
+ "type": "STATE",
45
+ "shape": [
46
+ 6
47
+ ]
48
+ },
49
+ "observation.state.right_arm": {
50
+ "type": "STATE",
51
+ "shape": [
52
+ 6
53
+ ]
54
+ },
55
+ "observation.state.right_arm.velocities": {
56
+ "type": "STATE",
57
+ "shape": [
58
+ 6
59
+ ]
60
+ },
61
+ "observation.state.chassis.imu": {
62
+ "type": "STATE",
63
+ "shape": [
64
+ 10
65
+ ]
66
+ },
67
+ "observation.state.chassis": {
68
+ "type": "STATE",
69
+ "shape": [
70
+ 3
71
+ ]
72
+ },
73
+ "observation.state.chassis.velocities": {
74
+ "type": "STATE",
75
+ "shape": [
76
+ 3
77
+ ]
78
+ },
79
+ "observation.state.torso": {
80
+ "type": "STATE",
81
+ "shape": [
82
+ 4
83
+ ]
84
+ },
85
+ "observation.state.torso.velocities": {
86
+ "type": "STATE",
87
+ "shape": [
88
+ 4
89
+ ]
90
+ },
91
+ "observation.state.left_gripper": {
92
+ "type": "STATE",
93
+ "shape": [
94
+ 1
95
+ ]
96
+ },
97
+ "observation.state.right_gripper": {
98
+ "type": "STATE",
99
+ "shape": [
100
+ 1
101
+ ]
102
+ },
103
+ "observation.state.left_ee_pose": {
104
+ "type": "STATE",
105
+ "shape": [
106
+ 7
107
+ ]
108
+ },
109
+ "observation.state.right_ee_pose": {
110
+ "type": "STATE",
111
+ "shape": [
112
+ 7
113
+ ]
114
+ },
115
+ "observation.state": {
116
+ "type": "STATE",
117
+ "shape": [
118
+ 14
119
+ ]
120
+ }
121
+ },
122
+ "output_features": {
123
+ "action.left_gripper": {
124
+ "type": "ACTION",
125
+ "shape": [
126
+ 1
127
+ ]
128
+ },
129
+ "action.right_gripper": {
130
+ "type": "ACTION",
131
+ "shape": [
132
+ 1
133
+ ]
134
+ },
135
+ "action.chassis.velocities": {
136
+ "type": "ACTION",
137
+ "shape": [
138
+ 6
139
+ ]
140
+ },
141
+ "action.torso.velocities": {
142
+ "type": "ACTION",
143
+ "shape": [
144
+ 6
145
+ ]
146
+ },
147
+ "action.left_arm": {
148
+ "type": "ACTION",
149
+ "shape": [
150
+ 6
151
+ ]
152
+ },
153
+ "action.right_arm": {
154
+ "type": "ACTION",
155
+ "shape": [
156
+ 6
157
+ ]
158
+ },
159
+ "action": {
160
+ "type": "ACTION",
161
+ "shape": [
162
+ 14
163
+ ]
164
+ }
165
+ },
166
+ "device": "cuda",
167
+ "use_amp": false,
168
+ "use_peft": false,
169
+ "push_to_hub": true,
170
+ "repo_id": "XYZPIT/multitask_dit_260413_r1lite_",
171
+ "private": null,
172
+ "tags": null,
173
+ "license": null,
174
+ "pretrained_path": null,
175
+ "horizon": 32,
176
+ "n_action_steps": 24,
177
+ "objective": "diffusion",
178
+ "noise_scheduler_type": "DDIM",
179
+ "num_train_timesteps": 10,
180
+ "beta_schedule": "squaredcos_cap_v2",
181
+ "beta_start": 0.0001,
182
+ "beta_end": 0.02,
183
+ "prediction_type": "epsilon",
184
+ "clip_sample": true,
185
+ "clip_sample_range": 1.0,
186
+ "num_inference_steps": null,
187
+ "sigma_min": 0.0,
188
+ "num_integration_steps": 10,
189
+ "integration_method": "euler",
190
+ "timestep_sampling_strategy": "beta",
191
+ "timestep_sampling_s": 0.999,
192
+ "timestep_sampling_alpha": 1.5,
193
+ "timestep_sampling_beta": 1.0,
194
+ "hidden_dim": 512,
195
+ "num_layers": 6,
196
+ "num_heads": 8,
197
+ "dropout": 0.1,
198
+ "use_positional_encoding": false,
199
+ "timestep_embed_dim": 256,
200
+ "use_rope": true,
201
+ "rope_base": 10000.0,
202
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
203
+ "use_separate_rgb_encoder_per_camera": false,
204
+ "vision_encoder_lr_multiplier": 0.1,
205
+ "image_resize_shape": [
206
+ 256,
207
+ 256
208
+ ],
209
+ "image_crop_shape": [
210
+ 224,
211
+ 224
212
+ ],
213
+ "image_crop_is_random": true,
214
+ "text_encoder_name": "openai/clip-vit-base-patch16",
215
+ "tokenizer_max_length": 77,
216
+ "tokenizer_padding": "max_length",
217
+ "tokenizer_padding_side": "right",
218
+ "tokenizer_truncation": true,
219
+ "normalization_mapping": {
220
+ "VISUAL": "MEAN_STD",
221
+ "STATE": "MIN_MAX",
222
+ "ACTION": "MIN_MAX"
223
+ },
224
+ "optimizer_lr": 2e-05,
225
+ "optimizer_betas": [
226
+ 0.95,
227
+ 0.999
228
+ ],
229
+ "optimizer_eps": 1e-08,
230
+ "optimizer_weight_decay": 0.0,
231
+ "scheduler_name": "cosine",
232
+ "scheduler_warmup_steps": 0,
233
+ "do_mask_loss_for_padding": false,
234
+ "drop_n_last_frames": 7
235
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:309275d1408597b7b7548675413d6f20526a61965d7292f524f4c440813b7cf7
3
+ size 1223199240
train_config.json ADDED
@@ -0,0 +1,374 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "XYZPIT/260413_r1lite_lerobot",
4
+ "root": "/home/xyz-ai/.cache/huggingface/hub/datasets--XYZPIT--260413_r1lite_lerobot/snapshots/419885cde5996f80240a0e4ad0e154e28e9ddc70",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "return_uint8": false,
81
+ "streaming": false
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "multi_task_dit",
86
+ "n_obs_steps": 2,
87
+ "input_features": {
88
+ "observation.images.head_rgb": {
89
+ "type": "VISUAL",
90
+ "shape": [
91
+ 3,
92
+ 720,
93
+ 1280
94
+ ]
95
+ },
96
+ "observation.images.head_right_rgb": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 3,
100
+ 720,
101
+ 1280
102
+ ]
103
+ },
104
+ "observation.images.left_wrist_rgb": {
105
+ "type": "VISUAL",
106
+ "shape": [
107
+ 3,
108
+ 360,
109
+ 640
110
+ ]
111
+ },
112
+ "observation.images.right_wrist_rgb": {
113
+ "type": "VISUAL",
114
+ "shape": [
115
+ 3,
116
+ 360,
117
+ 640
118
+ ]
119
+ },
120
+ "observation.state.left_arm": {
121
+ "type": "STATE",
122
+ "shape": [
123
+ 6
124
+ ]
125
+ },
126
+ "observation.state.left_arm.velocities": {
127
+ "type": "STATE",
128
+ "shape": [
129
+ 6
130
+ ]
131
+ },
132
+ "observation.state.right_arm": {
133
+ "type": "STATE",
134
+ "shape": [
135
+ 6
136
+ ]
137
+ },
138
+ "observation.state.right_arm.velocities": {
139
+ "type": "STATE",
140
+ "shape": [
141
+ 6
142
+ ]
143
+ },
144
+ "observation.state.chassis.imu": {
145
+ "type": "STATE",
146
+ "shape": [
147
+ 10
148
+ ]
149
+ },
150
+ "observation.state.chassis": {
151
+ "type": "STATE",
152
+ "shape": [
153
+ 3
154
+ ]
155
+ },
156
+ "observation.state.chassis.velocities": {
157
+ "type": "STATE",
158
+ "shape": [
159
+ 3
160
+ ]
161
+ },
162
+ "observation.state.torso": {
163
+ "type": "STATE",
164
+ "shape": [
165
+ 4
166
+ ]
167
+ },
168
+ "observation.state.torso.velocities": {
169
+ "type": "STATE",
170
+ "shape": [
171
+ 4
172
+ ]
173
+ },
174
+ "observation.state.left_gripper": {
175
+ "type": "STATE",
176
+ "shape": [
177
+ 1
178
+ ]
179
+ },
180
+ "observation.state.right_gripper": {
181
+ "type": "STATE",
182
+ "shape": [
183
+ 1
184
+ ]
185
+ },
186
+ "observation.state.left_ee_pose": {
187
+ "type": "STATE",
188
+ "shape": [
189
+ 7
190
+ ]
191
+ },
192
+ "observation.state.right_ee_pose": {
193
+ "type": "STATE",
194
+ "shape": [
195
+ 7
196
+ ]
197
+ },
198
+ "observation.state": {
199
+ "type": "STATE",
200
+ "shape": [
201
+ 14
202
+ ]
203
+ }
204
+ },
205
+ "output_features": {
206
+ "action.left_gripper": {
207
+ "type": "ACTION",
208
+ "shape": [
209
+ 1
210
+ ]
211
+ },
212
+ "action.right_gripper": {
213
+ "type": "ACTION",
214
+ "shape": [
215
+ 1
216
+ ]
217
+ },
218
+ "action.chassis.velocities": {
219
+ "type": "ACTION",
220
+ "shape": [
221
+ 6
222
+ ]
223
+ },
224
+ "action.torso.velocities": {
225
+ "type": "ACTION",
226
+ "shape": [
227
+ 6
228
+ ]
229
+ },
230
+ "action.left_arm": {
231
+ "type": "ACTION",
232
+ "shape": [
233
+ 6
234
+ ]
235
+ },
236
+ "action.right_arm": {
237
+ "type": "ACTION",
238
+ "shape": [
239
+ 6
240
+ ]
241
+ },
242
+ "action": {
243
+ "type": "ACTION",
244
+ "shape": [
245
+ 14
246
+ ]
247
+ }
248
+ },
249
+ "device": "cuda",
250
+ "use_amp": false,
251
+ "use_peft": false,
252
+ "push_to_hub": true,
253
+ "repo_id": "XYZPIT/multitask_dit_260413_r1lite_",
254
+ "private": null,
255
+ "tags": null,
256
+ "license": null,
257
+ "pretrained_path": null,
258
+ "horizon": 32,
259
+ "n_action_steps": 24,
260
+ "objective": "diffusion",
261
+ "noise_scheduler_type": "DDIM",
262
+ "num_train_timesteps": 10,
263
+ "beta_schedule": "squaredcos_cap_v2",
264
+ "beta_start": 0.0001,
265
+ "beta_end": 0.02,
266
+ "prediction_type": "epsilon",
267
+ "clip_sample": true,
268
+ "clip_sample_range": 1.0,
269
+ "num_inference_steps": null,
270
+ "sigma_min": 0.0,
271
+ "num_integration_steps": 10,
272
+ "integration_method": "euler",
273
+ "timestep_sampling_strategy": "beta",
274
+ "timestep_sampling_s": 0.999,
275
+ "timestep_sampling_alpha": 1.5,
276
+ "timestep_sampling_beta": 1.0,
277
+ "hidden_dim": 512,
278
+ "num_layers": 6,
279
+ "num_heads": 8,
280
+ "dropout": 0.1,
281
+ "use_positional_encoding": false,
282
+ "timestep_embed_dim": 256,
283
+ "use_rope": true,
284
+ "rope_base": 10000.0,
285
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
286
+ "use_separate_rgb_encoder_per_camera": false,
287
+ "vision_encoder_lr_multiplier": 0.1,
288
+ "image_resize_shape": [
289
+ 256,
290
+ 256
291
+ ],
292
+ "image_crop_shape": [
293
+ 224,
294
+ 224
295
+ ],
296
+ "image_crop_is_random": true,
297
+ "text_encoder_name": "openai/clip-vit-base-patch16",
298
+ "tokenizer_max_length": 77,
299
+ "tokenizer_padding": "max_length",
300
+ "tokenizer_padding_side": "right",
301
+ "tokenizer_truncation": true,
302
+ "normalization_mapping": {
303
+ "VISUAL": "MEAN_STD",
304
+ "STATE": "MIN_MAX",
305
+ "ACTION": "MIN_MAX"
306
+ },
307
+ "optimizer_lr": 2e-05,
308
+ "optimizer_betas": [
309
+ 0.95,
310
+ 0.999
311
+ ],
312
+ "optimizer_eps": 1e-08,
313
+ "optimizer_weight_decay": 0.0,
314
+ "scheduler_name": "cosine",
315
+ "scheduler_warmup_steps": 0,
316
+ "do_mask_loss_for_padding": false,
317
+ "drop_n_last_frames": 7
318
+ },
319
+ "output_dir": "outputs/mutitask_dit_training",
320
+ "job_name": "multi_task_dit",
321
+ "resume": false,
322
+ "seed": 1000,
323
+ "cudnn_deterministic": false,
324
+ "num_workers": 4,
325
+ "batch_size": 32,
326
+ "prefetch_factor": 4,
327
+ "persistent_workers": true,
328
+ "steps": 100000,
329
+ "eval_freq": 20000,
330
+ "log_freq": 500,
331
+ "tolerance_s": 0.0001,
332
+ "save_checkpoint": true,
333
+ "save_freq": 20000,
334
+ "use_policy_training_preset": true,
335
+ "optimizer": {
336
+ "type": "adam",
337
+ "lr": 2e-05,
338
+ "weight_decay": 0.0,
339
+ "grad_clip_norm": 10.0,
340
+ "betas": [
341
+ 0.95,
342
+ 0.999
343
+ ],
344
+ "eps": 1e-08
345
+ },
346
+ "scheduler": {
347
+ "type": "diffuser",
348
+ "num_warmup_steps": 0,
349
+ "name": "cosine"
350
+ },
351
+ "eval": {
352
+ "n_episodes": 50,
353
+ "batch_size": 14,
354
+ "use_async_envs": true
355
+ },
356
+ "wandb": {
357
+ "enable": true,
358
+ "disable_artifact": false,
359
+ "project": "lerobot",
360
+ "entity": null,
361
+ "notes": null,
362
+ "run_id": "exbr3cz8",
363
+ "mode": null,
364
+ "add_tags": true
365
+ },
366
+ "peft": null,
367
+ "use_rabc": false,
368
+ "rabc_progress_path": null,
369
+ "rabc_kappa": 0.01,
370
+ "rabc_epsilon": 1e-06,
371
+ "rabc_head_mode": "sparse",
372
+ "rename_map": {},
373
+ "checkpoint_path": null
374
+ }