Hugging Face's logo

TeleEmbodied
/
VISTA

Safetensors
VLA
Model card Files
xet
 Community
tgzzglz666 commited on
Commit
aa7ee07
·
verified ·
1 Parent(s): d720849

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. pretrained_model/config.json +84 -0
  2. pretrained_model/model.safetensors +3 -0
  3. pretrained_model/policy_postprocessor.json +32 -0
  4. pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. pretrained_model/policy_preprocessor.json +81 -0
  6. pretrained_model/policy_preprocessor_step_2_normalizer_processor.safetensors +3 -0
  7. pretrained_model/train_config.json +276 -0
pretrained_model/config.json ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "vista",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 16
9
+ ]
10
+ },
11
+ "observation.images.left_wrist": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 224,
16
+ 224
17
+ ]
18
+ },
19
+ "observation.images.right_wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 224,
24
+ 224
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 16
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "use_delta_action": true,
39
+ "push_to_hub": false,
40
+ "repo_id": null,
41
+ "private": null,
42
+ "tags": null,
43
+ "license": null,
44
+ "pretrained_path": "/data/ysy/data/models/umi_0312_40w",
45
+ "paligemma_variant": "gemma_2b",
46
+ "action_expert_variant": "gemma_300m",
47
+ "dtype": "bfloat16",
48
+ "chunk_size": 50,
49
+ "n_action_steps": 50,
50
+ "max_state_dim": 32,
51
+ "max_action_dim": 32,
52
+ "num_inference_steps": 10,
53
+ "time_sampling_beta_alpha": 1.5,
54
+ "time_sampling_beta_beta": 1.0,
55
+ "time_sampling_scale": 0.999,
56
+ "time_sampling_offset": 0.001,
57
+ "min_period": 0.004,
58
+ "max_period": 4.0,
59
+ "image_resolution": [
60
+ 224,
61
+ 224
62
+ ],
63
+ "empty_cameras": 0,
64
+ "tokenizer_max_length": 200,
65
+ "normalization_mapping": {
66
+ "VISUAL": "IDENTITY",
67
+ "STATE": "QUANTILES",
68
+ "ACTION": "QUANTILES"
69
+ },
70
+ "gradient_checkpointing": false,
71
+ "compile_model": false,
72
+ "compile_mode": "max-autotune",
73
+ "optimizer_lr": 2.5e-05,
74
+ "optimizer_betas": [
75
+ 0.9,
76
+ 0.95
77
+ ],
78
+ "optimizer_eps": 1e-08,
79
+ "optimizer_weight_decay": 0.01,
80
+ "optimizer_grad_clip_norm": 1.0,
81
+ "scheduler_warmup_steps": 1000,
82
+ "scheduler_decay_steps": 300000,
83
+ "scheduler_decay_lr": 2.5e-06
84
+ }
pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b92361ce0150cf73773b6b8c966937942435aa02c8ed541f3ea3bd78edd1949
3
+ size 7473096344
pretrained_model/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
+ 16
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "IDENTITY",
18
+ "STATE": "QUANTILES",
19
+ "ACTION": "QUANTILES"
20
+ }
21
+ },
22
+ "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
23
+ },
24
+ {
25
+ "registry_name": "device_processor",
26
+ "config": {
27
+ "device": "cuda",
28
+ "float_dtype": null
29
+ }
30
+ }
31
+ ]
32
+ }
pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe9b5c1fd200eb6ea91c2adbe2c77b2d6acd425636ef033c9a99b98443b7be6f
3
+ size 6748
pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "normalizer_processor",
16
+ "config": {
17
+ "eps": 1e-08,
18
+ "features": {
19
+ "observation.state": {
20
+ "type": "STATE",
21
+ "shape": [
22
+ 16
23
+ ]
24
+ },
25
+ "observation.images.robot_0": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 224,
30
+ 224
31
+ ]
32
+ },
33
+ "observation.images.robot_1": {
34
+ "type": "VISUAL",
35
+ "shape": [
36
+ 3,
37
+ 224,
38
+ 224
39
+ ]
40
+ },
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 16
45
+ ]
46
+ }
47
+ },
48
+ "norm_map": {
49
+ "VISUAL": "IDENTITY",
50
+ "STATE": "QUANTILES",
51
+ "ACTION": "QUANTILES"
52
+ }
53
+ },
54
+ "state_file": "policy_preprocessor_step_2_normalizer_processor.safetensors"
55
+ },
56
+ {
57
+ "registry_name": "vista_prepare_state_tokenizer_processor_step",
58
+ "config": {
59
+ "max_state_dim": 16
60
+ }
61
+ },
62
+ {
63
+ "registry_name": "tokenizer_processor",
64
+ "config": {
65
+ "max_length": 200,
66
+ "task_key": "task",
67
+ "padding_side": "right",
68
+ "padding": "max_length",
69
+ "truncation": true,
70
+ "tokenizer_name": "/data/ysy/data/models/plgm3b"
71
+ }
72
+ },
73
+ {
74
+ "registry_name": "device_processor",
75
+ "config": {
76
+ "device": "cuda",
77
+ "float_dtype": null
78
+ }
79
+ }
80
+ ]
81
+ }
pretrained_model/policy_preprocessor_step_2_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fe9b5c1fd200eb6ea91c2adbe2c77b2d6acd425636ef033c9a99b98443b7be6f
3
+ size 6748
pretrained_model/train_config.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": null,
4
+ "root": "/data/ysy/data/lerobot_pretrain_used",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
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
+ "wrist_transforms": {
78
+ "enable": true,
79
+ "max_num_transforms": 3,
80
+ "random_order": false,
81
+ "tfs": {
82
+ "brightness": {
83
+ "weight": 1.0,
84
+ "type": "ColorJitter",
85
+ "kwargs": {
86
+ "brightness": [
87
+ 0.8,
88
+ 1.2
89
+ ]
90
+ }
91
+ },
92
+ "contrast": {
93
+ "weight": 1.0,
94
+ "type": "ColorJitter",
95
+ "kwargs": {
96
+ "contrast": [
97
+ 0.8,
98
+ 1.2
99
+ ]
100
+ }
101
+ },
102
+ "saturation": {
103
+ "weight": 1.0,
104
+ "type": "ColorJitter",
105
+ "kwargs": {
106
+ "saturation": [
107
+ 0.5,
108
+ 1.5
109
+ ]
110
+ }
111
+ },
112
+ "hue": {
113
+ "weight": 1.0,
114
+ "type": "ColorJitter",
115
+ "kwargs": {
116
+ "hue": [
117
+ -0.05,
118
+ 0.05
119
+ ]
120
+ }
121
+ },
122
+ "sharpness": {
123
+ "weight": 1.0,
124
+ "type": "SharpnessJitter",
125
+ "kwargs": {
126
+ "sharpness": [
127
+ 0.5,
128
+ 1.5
129
+ ]
130
+ }
131
+ }
132
+ }
133
+ },
134
+ "revision": "v2.0",
135
+ "use_imagenet_stats": true,
136
+ "video_backend": "torchcodec",
137
+ "streaming": false
138
+ },
139
+ "env": null,
140
+ "policy": {
141
+ "type": "vista",
142
+ "n_obs_steps": 1,
143
+ "input_features": {
144
+ "observation.state": {
145
+ "type": "STATE",
146
+ "shape": [
147
+ 16
148
+ ]
149
+ },
150
+ "observation.images.left_wrist": {
151
+ "type": "VISUAL",
152
+ "shape": [
153
+ 3,
154
+ 224,
155
+ 224
156
+ ]
157
+ },
158
+ "observation.images.right_wrist": {
159
+ "type": "VISUAL",
160
+ "shape": [
161
+ 3,
162
+ 224,
163
+ 224
164
+ ]
165
+ }
166
+ },
167
+ "output_features": {
168
+ "action": {
169
+ "type": "ACTION",
170
+ "shape": [
171
+ 16
172
+ ]
173
+ }
174
+ },
175
+ "device": "cuda",
176
+ "use_amp": false,
177
+ "use_delta_action": true,
178
+ "push_to_hub": false,
179
+ "repo_id": null,
180
+ "private": null,
181
+ "tags": null,
182
+ "license": null,
183
+ "pretrained_path": "/data/ysy/data/models/umi_0312_40w",
184
+ "paligemma_variant": "gemma_2b",
185
+ "action_expert_variant": "gemma_300m",
186
+ "dtype": "bfloat16",
187
+ "chunk_size": 50,
188
+ "n_action_steps": 50,
189
+ "max_state_dim": 32,
190
+ "max_action_dim": 32,
191
+ "num_inference_steps": 10,
192
+ "time_sampling_beta_alpha": 1.5,
193
+ "time_sampling_beta_beta": 1.0,
194
+ "time_sampling_scale": 0.999,
195
+ "time_sampling_offset": 0.001,
196
+ "min_period": 0.004,
197
+ "max_period": 4.0,
198
+ "image_resolution": [
199
+ 224,
200
+ 224
201
+ ],
202
+ "empty_cameras": 0,
203
+ "tokenizer_max_length": 200,
204
+ "normalization_mapping": {
205
+ "VISUAL": "IDENTITY",
206
+ "STATE": "QUANTILES",
207
+ "ACTION": "QUANTILES"
208
+ },
209
+ "gradient_checkpointing": false,
210
+ "compile_model": false,
211
+ "compile_mode": "max-autotune",
212
+ "optimizer_lr": 2.5e-05,
213
+ "optimizer_betas": [
214
+ 0.9,
215
+ 0.95
216
+ ],
217
+ "optimizer_eps": 1e-08,
218
+ "optimizer_weight_decay": 0.01,
219
+ "optimizer_grad_clip_norm": 1.0,
220
+ "scheduler_warmup_steps": 1000,
221
+ "scheduler_decay_steps": 300000,
222
+ "scheduler_decay_lr": 2.5e-06
223
+ },
224
+ "output_dir": "/data/ysy/data/train_output/pre_s2/lerobot_pretrain_used/26-04-23_04-23-37_pi05_gpu8_ck50_lr2_5e-5_bs24_s300K_seed42",
225
+ "job_name": "vista",
226
+ "resume": false,
227
+ "seed": 42,
228
+ "num_workers": 8,
229
+ "batch_size": 24,
230
+ "steps": 300000,
231
+ "eval_freq": 5000,
232
+ "log_freq": 50,
233
+ "save_checkpoint": true,
234
+ "save_freq": 20000,
235
+ "use_policy_training_preset": true,
236
+ "optimizer": {
237
+ "type": "adamw",
238
+ "lr": 2.5e-05,
239
+ "weight_decay": 0.01,
240
+ "grad_clip_norm": 1.0,
241
+ "betas": [
242
+ 0.9,
243
+ 0.95
244
+ ],
245
+ "eps": 1e-08
246
+ },
247
+ "scheduler": {
248
+ "type": "cosine_decay_with_warmup",
249
+ "num_warmup_steps": 1000,
250
+ "num_decay_steps": 300000,
251
+ "peak_lr": 2.5e-05,
252
+ "decay_lr": 2.5e-06
253
+ },
254
+ "eval": {
255
+ "n_episodes": 50,
256
+ "batch_size": 50,
257
+ "use_async_envs": false
258
+ },
259
+ "wandb": {
260
+ "enable": false,
261
+ "disable_artifact": false,
262
+ "project": "lerobot",
263
+ "entity": null,
264
+ "notes": null,
265
+ "run_id": null,
266
+ "mode": null
267
+ },
268
+ "checkpoint_path": null,
269
+ "rename_map": {},
270
+ "accelerator_logging_dir": "accelerate_logs",
271
+ "mixed_precision": "bf16",
272
+ "enforce_input_output_replace": true,
273
+ "gradient_checkpointing": false,
274
+ "save_as_safetensors": true,
275
+ "load_from_safetensors": true
276
+ }