Gen1113 commited on
Commit
1f862e0
·
verified ·
1 Parent(s): b49ea37

Upload folder using huggingface_hub

Browse files
config.json ADDED
@@ -0,0 +1,142 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.image": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 256
11
+ ]
12
+ },
13
+ "observation.images.image2": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 256
19
+ ]
20
+ },
21
+ "observation.state": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 8
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 7
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": true,
38
+ "use_peft": false,
39
+ "push_to_hub": false,
40
+ "repo_id": null,
41
+ "private": null,
42
+ "tags": null,
43
+ "license": null,
44
+ "pretrained_path": null,
45
+ "horizon": 32,
46
+ "n_action_steps": 10,
47
+ "objective": "diffusion",
48
+ "noise_scheduler_type": "DDPM",
49
+ "num_train_timesteps": 100,
50
+ "beta_schedule": "squaredcos_cap_v2",
51
+ "beta_start": 0.0001,
52
+ "beta_end": 0.02,
53
+ "prediction_type": "epsilon",
54
+ "clip_sample": true,
55
+ "clip_sample_range": 1.0,
56
+ "num_inference_steps": null,
57
+ "sigma_min": 0.0,
58
+ "num_integration_steps": 100,
59
+ "integration_method": "euler",
60
+ "timestep_sampling_strategy": "beta",
61
+ "timestep_sampling_s": 0.999,
62
+ "timestep_sampling_alpha": 1.5,
63
+ "timestep_sampling_beta": 1.0,
64
+ "hidden_dim": 512,
65
+ "num_layers": 8,
66
+ "num_heads": 8,
67
+ "dropout": 0.1,
68
+ "modality_dropout_prob": 0.1,
69
+ "modality_weight_entropy_reg_weight": 0.0,
70
+ "modality_weight_balance_reg_weight": 0.0,
71
+ "use_positional_encoding": false,
72
+ "timestep_embed_dim": 128,
73
+ "use_rope": true,
74
+ "rope_base": 10000.0,
75
+ "qformer_num_queries": 8,
76
+ "qformer_hidden_dim": 512,
77
+ "qformer_num_layers": 2,
78
+ "qformer_num_heads": 8,
79
+ "qformer_dropout": 0.1,
80
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
81
+ "use_separate_rgb_encoder_per_camera": false,
82
+ "vision_encoder_lr_multiplier": 0.1,
83
+ "image_resize_shape": [
84
+ 256,
85
+ 256
86
+ ],
87
+ "image_crop_shape": [
88
+ 224,
89
+ 224
90
+ ],
91
+ "image_crop_is_random": true,
92
+ "robot_state_embed_dim": 128,
93
+ "text_embed_dim": 128,
94
+ "text_encoder_name": "openai/clip-vit-base-patch16",
95
+ "tokenizer_max_length": 77,
96
+ "tokenizer_padding": "max_length",
97
+ "tokenizer_padding_side": "right",
98
+ "tokenizer_truncation": true,
99
+ "normalization_mapping": {
100
+ "VISUAL": "MEAN_STD",
101
+ "STATE": "MIN_MAX",
102
+ "ACTION": "MIN_MAX"
103
+ },
104
+ "optimizer_lr": 0.0003,
105
+ "optimizer_betas": [
106
+ 0.95,
107
+ 0.999
108
+ ],
109
+ "optimizer_eps": 1e-08,
110
+ "optimizer_weight_decay": 1e-06,
111
+ "scheduler_name": "cosine",
112
+ "scheduler_warmup_steps": 500,
113
+ "do_mask_loss_for_padding": false,
114
+ "robust_adapter_training": false,
115
+ "robust_adapter_hidden_dim": 512,
116
+ "robust_adapter_num_queries": 4,
117
+ "robust_adapter_dropout": 0.0,
118
+ "robust_feature_loss_weight": 1.0,
119
+ "robust_action_loss_weight": 1.0,
120
+ "robust_clean_action_loss_weight": 0.0,
121
+ "robust_noise_std": 0.08,
122
+ "robust_patch_noise_prob": 0.5,
123
+ "robust_patch_noise_ratio": [
124
+ 0.12,
125
+ 0.35
126
+ ],
127
+ "robust_high_freq_prob": 0.5,
128
+ "robust_photometric_prob": 0.5,
129
+ "robust_brightness": 0.4,
130
+ "robust_contrast": 0.5,
131
+ "robust_saturation": 0.6,
132
+ "robust_hue": 0.15,
133
+ "robust_affine_prob": 0.5,
134
+ "robust_affine_max_degrees": 15.0,
135
+ "robust_affine_max_translate": 0.1,
136
+ "robust_affine_scale": [
137
+ 0.9,
138
+ 1.1
139
+ ],
140
+ "robust_vertical_flip_prob": 0.25,
141
+ "drop_n_last_frames": 21
142
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9efbd63cd5b842c0a07c0d6fbe71ff167aa52ac63e12b8a4742e25e8e1a6e5aa
3
+ size 1049795820
policy_postprocessor.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 7
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "MEAN_STD",
18
+ "STATE": "MIN_MAX",
19
+ "ACTION": "MIN_MAX"
20
+ }
21
+ },
22
+ "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
23
+ },
24
+ {
25
+ "registry_name": "device_processor",
26
+ "config": {
27
+ "device": "cpu",
28
+ "float_dtype": null
29
+ }
30
+ }
31
+ ]
32
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a002c0df7f79c5b169c5a899ad151d4ea1bed246c7d82bd93ed1556558d517a9
3
+ size 7668
policy_preprocessor.json ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "tokenizer_processor",
16
+ "config": {
17
+ "max_length": 77,
18
+ "task_key": "task",
19
+ "padding_side": "right",
20
+ "padding": "max_length",
21
+ "truncation": true,
22
+ "tokenizer_name": "openai/clip-vit-base-patch16"
23
+ }
24
+ },
25
+ {
26
+ "registry_name": "device_processor",
27
+ "config": {
28
+ "device": "cuda",
29
+ "float_dtype": null
30
+ }
31
+ },
32
+ {
33
+ "registry_name": "normalizer_processor",
34
+ "config": {
35
+ "eps": 1e-08,
36
+ "features": {
37
+ "observation.images.image": {
38
+ "type": "VISUAL",
39
+ "shape": [
40
+ 3,
41
+ 256,
42
+ 256
43
+ ]
44
+ },
45
+ "observation.images.image2": {
46
+ "type": "VISUAL",
47
+ "shape": [
48
+ 3,
49
+ 256,
50
+ 256
51
+ ]
52
+ },
53
+ "observation.state": {
54
+ "type": "STATE",
55
+ "shape": [
56
+ 8
57
+ ]
58
+ },
59
+ "action": {
60
+ "type": "ACTION",
61
+ "shape": [
62
+ 7
63
+ ]
64
+ }
65
+ },
66
+ "norm_map": {
67
+ "VISUAL": "MEAN_STD",
68
+ "STATE": "MIN_MAX",
69
+ "ACTION": "MIN_MAX"
70
+ }
71
+ },
72
+ "state_file": "policy_preprocessor_step_4_normalizer_processor.safetensors"
73
+ }
74
+ ]
75
+ }
policy_preprocessor_step_4_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a002c0df7f79c5b169c5a899ad151d4ea1bed246c7d82bd93ed1556558d517a9
3
+ size 7668
train_config.json ADDED
@@ -0,0 +1,373 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "HuggingFaceVLA/libero",
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
+ },
82
+ "env": {
83
+ "type": "libero",
84
+ "task": "libero_object",
85
+ "fps": 30,
86
+ "features": {
87
+ "action": {
88
+ "type": "ACTION",
89
+ "shape": [
90
+ 7
91
+ ]
92
+ },
93
+ "pixels/agentview_image": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 360,
97
+ 360,
98
+ 3
99
+ ]
100
+ },
101
+ "pixels/robot0_eye_in_hand_image": {
102
+ "type": "VISUAL",
103
+ "shape": [
104
+ 360,
105
+ 360,
106
+ 3
107
+ ]
108
+ },
109
+ "robot_state/eef/pos": {
110
+ "type": "STATE",
111
+ "shape": [
112
+ 3
113
+ ]
114
+ },
115
+ "robot_state/eef/quat": {
116
+ "type": "STATE",
117
+ "shape": [
118
+ 4
119
+ ]
120
+ },
121
+ "robot_state/eef/mat": {
122
+ "type": "STATE",
123
+ "shape": [
124
+ 3,
125
+ 3
126
+ ]
127
+ },
128
+ "robot_state/gripper/qpos": {
129
+ "type": "STATE",
130
+ "shape": [
131
+ 2
132
+ ]
133
+ },
134
+ "robot_state/gripper/qvel": {
135
+ "type": "STATE",
136
+ "shape": [
137
+ 2
138
+ ]
139
+ },
140
+ "robot_state/joints/pos": {
141
+ "type": "STATE",
142
+ "shape": [
143
+ 7
144
+ ]
145
+ },
146
+ "robot_state/joints/vel": {
147
+ "type": "STATE",
148
+ "shape": [
149
+ 7
150
+ ]
151
+ }
152
+ },
153
+ "features_map": {
154
+ "action": "action",
155
+ "robot_state/eef/pos": "observation.state.eef_pos",
156
+ "robot_state/eef/quat": "observation.state.eef_quat",
157
+ "robot_state/eef/mat": "observation.state.eef_mat",
158
+ "robot_state/gripper/qpos": "observation.state.gripper_qpos",
159
+ "robot_state/gripper/qvel": "observation.state.gripper_qvel",
160
+ "robot_state/joints/pos": "observation.state.joint_pos",
161
+ "robot_state/joints/vel": "observation.state.joint_vel",
162
+ "pixels/agentview_image": "observation.images.image",
163
+ "pixels/robot0_eye_in_hand_image": "observation.images.image2"
164
+ },
165
+ "max_parallel_tasks": 1,
166
+ "disable_env_checker": true,
167
+ "task_ids": null,
168
+ "episode_length": null,
169
+ "obs_type": "pixels_agent_pos",
170
+ "render_mode": "rgb_array",
171
+ "camera_name": "agentview_image,robot0_eye_in_hand_image",
172
+ "init_states": true,
173
+ "camera_name_mapping": null,
174
+ "observation_height": 360,
175
+ "observation_width": 360,
176
+ "control_mode": "relative"
177
+ },
178
+ "policy": {
179
+ "type": "multi_task_dit",
180
+ "n_obs_steps": 2,
181
+ "input_features": {
182
+ "observation.images.image": {
183
+ "type": "VISUAL",
184
+ "shape": [
185
+ 3,
186
+ 256,
187
+ 256
188
+ ]
189
+ },
190
+ "observation.images.image2": {
191
+ "type": "VISUAL",
192
+ "shape": [
193
+ 3,
194
+ 256,
195
+ 256
196
+ ]
197
+ },
198
+ "observation.state": {
199
+ "type": "STATE",
200
+ "shape": [
201
+ 8
202
+ ]
203
+ }
204
+ },
205
+ "output_features": {
206
+ "action": {
207
+ "type": "ACTION",
208
+ "shape": [
209
+ 7
210
+ ]
211
+ }
212
+ },
213
+ "device": "cuda",
214
+ "use_amp": true,
215
+ "use_peft": false,
216
+ "push_to_hub": false,
217
+ "repo_id": null,
218
+ "private": null,
219
+ "tags": null,
220
+ "license": null,
221
+ "pretrained_path": null,
222
+ "horizon": 32,
223
+ "n_action_steps": 10,
224
+ "objective": "diffusion",
225
+ "noise_scheduler_type": "DDPM",
226
+ "num_train_timesteps": 100,
227
+ "beta_schedule": "squaredcos_cap_v2",
228
+ "beta_start": 0.0001,
229
+ "beta_end": 0.02,
230
+ "prediction_type": "epsilon",
231
+ "clip_sample": true,
232
+ "clip_sample_range": 1.0,
233
+ "num_inference_steps": null,
234
+ "sigma_min": 0.0,
235
+ "num_integration_steps": 100,
236
+ "integration_method": "euler",
237
+ "timestep_sampling_strategy": "beta",
238
+ "timestep_sampling_s": 0.999,
239
+ "timestep_sampling_alpha": 1.5,
240
+ "timestep_sampling_beta": 1.0,
241
+ "hidden_dim": 512,
242
+ "num_layers": 8,
243
+ "num_heads": 8,
244
+ "dropout": 0.1,
245
+ "modality_dropout_prob": 0.1,
246
+ "modality_weight_entropy_reg_weight": 0.0,
247
+ "modality_weight_balance_reg_weight": 0.0,
248
+ "use_positional_encoding": false,
249
+ "timestep_embed_dim": 128,
250
+ "use_rope": true,
251
+ "rope_base": 10000.0,
252
+ "qformer_num_queries": 8,
253
+ "qformer_hidden_dim": 512,
254
+ "qformer_num_layers": 2,
255
+ "qformer_num_heads": 8,
256
+ "qformer_dropout": 0.1,
257
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
258
+ "use_separate_rgb_encoder_per_camera": false,
259
+ "vision_encoder_lr_multiplier": 0.1,
260
+ "image_resize_shape": [
261
+ 256,
262
+ 256
263
+ ],
264
+ "image_crop_shape": [
265
+ 224,
266
+ 224
267
+ ],
268
+ "image_crop_is_random": true,
269
+ "robot_state_embed_dim": 128,
270
+ "text_embed_dim": 128,
271
+ "text_encoder_name": "openai/clip-vit-base-patch16",
272
+ "tokenizer_max_length": 77,
273
+ "tokenizer_padding": "max_length",
274
+ "tokenizer_padding_side": "right",
275
+ "tokenizer_truncation": true,
276
+ "normalization_mapping": {
277
+ "VISUAL": "MEAN_STD",
278
+ "STATE": "MIN_MAX",
279
+ "ACTION": "MIN_MAX"
280
+ },
281
+ "optimizer_lr": 0.0003,
282
+ "optimizer_betas": [
283
+ 0.95,
284
+ 0.999
285
+ ],
286
+ "optimizer_eps": 1e-08,
287
+ "optimizer_weight_decay": 1e-06,
288
+ "scheduler_name": "cosine",
289
+ "scheduler_warmup_steps": 500,
290
+ "do_mask_loss_for_padding": false,
291
+ "robust_adapter_training": false,
292
+ "robust_adapter_hidden_dim": 512,
293
+ "robust_adapter_num_queries": 4,
294
+ "robust_adapter_dropout": 0.0,
295
+ "robust_feature_loss_weight": 1.0,
296
+ "robust_action_loss_weight": 1.0,
297
+ "robust_clean_action_loss_weight": 0.0,
298
+ "robust_noise_std": 0.08,
299
+ "robust_patch_noise_prob": 0.5,
300
+ "robust_patch_noise_ratio": [
301
+ 0.12,
302
+ 0.35
303
+ ],
304
+ "robust_high_freq_prob": 0.5,
305
+ "robust_photometric_prob": 0.5,
306
+ "robust_brightness": 0.4,
307
+ "robust_contrast": 0.5,
308
+ "robust_saturation": 0.6,
309
+ "robust_hue": 0.15,
310
+ "robust_affine_prob": 0.5,
311
+ "robust_affine_max_degrees": 15.0,
312
+ "robust_affine_max_translate": 0.1,
313
+ "robust_affine_scale": [
314
+ 0.9,
315
+ 1.1
316
+ ],
317
+ "robust_vertical_flip_prob": 0.25,
318
+ "drop_n_last_frames": 21
319
+ },
320
+ "output_dir": "outputs/multi_task_dit/libero-object/multi_task_dit-bs64-asteps10-40k",
321
+ "job_name": "multi_task_dit-bs64-asteps10-40k",
322
+ "resume": false,
323
+ "seed": 1000,
324
+ "cudnn_deterministic": false,
325
+ "num_workers": 12,
326
+ "batch_size": 64,
327
+ "steps": 40000,
328
+ "eval_freq": 40001,
329
+ "log_freq": 100,
330
+ "tolerance_s": 0.0001,
331
+ "save_checkpoint": true,
332
+ "save_freq": 40000,
333
+ "use_policy_training_preset": true,
334
+ "optimizer": {
335
+ "type": "adam",
336
+ "lr": 0.0003,
337
+ "weight_decay": 1e-06,
338
+ "grad_clip_norm": 10.0,
339
+ "betas": [
340
+ 0.95,
341
+ 0.999
342
+ ],
343
+ "eps": 1e-08
344
+ },
345
+ "scheduler": {
346
+ "type": "diffuser",
347
+ "num_warmup_steps": 500,
348
+ "name": "cosine"
349
+ },
350
+ "eval": {
351
+ "n_episodes": 10,
352
+ "batch_size": 1,
353
+ "use_async_envs": false
354
+ },
355
+ "wandb": {
356
+ "enable": true,
357
+ "disable_artifact": false,
358
+ "project": "libero-object",
359
+ "entity": "li-gen",
360
+ "notes": null,
361
+ "run_id": "multi_task_dit-bs64-asteps10-40k",
362
+ "mode": "offline",
363
+ "add_tags": true
364
+ },
365
+ "peft": null,
366
+ "use_rabc": false,
367
+ "rabc_progress_path": null,
368
+ "rabc_kappa": 0.01,
369
+ "rabc_epsilon": 1e-06,
370
+ "rabc_head_mode": "sparse",
371
+ "rename_map": {},
372
+ "checkpoint_path": null
373
+ }