Hugging Face's logo

ywu67
/
pi05-task0000

Safetensors
Model card Files
xet
Community
ywu67 commited on
Commit
d708fd6
·
verified ·
1 Parent(s): 3538cc7

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. 5000/config.json +151 -0
  2. 5000/model.safetensors +3 -0
  3. 5000/policy_postprocessor.json +32 -0
  4. 5000/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. 5000/policy_preprocessor.json +147 -0
  6. 5000/policy_preprocessor_step_2_normalizer_processor.safetensors +3 -0
  7. 5000/train_config.json +280 -0
5000/config.json ADDED
@@ -0,0 +1,151 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "pi05",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.rgb.left_wrist": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 480,
9
+ 480,
10
+ 3
11
+ ]
12
+ },
13
+ "observation.images.rgb.right_wrist": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 480,
17
+ 480,
18
+ 3
19
+ ]
20
+ },
21
+ "observation.images.rgb.head": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 720,
25
+ 720,
26
+ 3
27
+ ]
28
+ },
29
+ "observation.images.depth.left_wrist": {
30
+ "type": "VISUAL",
31
+ "shape": [
32
+ 480,
33
+ 480,
34
+ 3
35
+ ]
36
+ },
37
+ "observation.images.depth.right_wrist": {
38
+ "type": "VISUAL",
39
+ "shape": [
40
+ 480,
41
+ 480,
42
+ 3
43
+ ]
44
+ },
45
+ "observation.images.depth.head": {
46
+ "type": "VISUAL",
47
+ "shape": [
48
+ 720,
49
+ 720,
50
+ 3
51
+ ]
52
+ },
53
+ "observation.images.seg_instance_id.left_wrist": {
54
+ "type": "VISUAL",
55
+ "shape": [
56
+ 480,
57
+ 480,
58
+ 3
59
+ ]
60
+ },
61
+ "observation.images.seg_instance_id.right_wrist": {
62
+ "type": "VISUAL",
63
+ "shape": [
64
+ 480,
65
+ 480,
66
+ 3
67
+ ]
68
+ },
69
+ "observation.images.seg_instance_id.head": {
70
+ "type": "VISUAL",
71
+ "shape": [
72
+ 720,
73
+ 720,
74
+ 3
75
+ ]
76
+ },
77
+ "observation.cam_rel_poses": {
78
+ "type": "STATE",
79
+ "shape": [
80
+ 21
81
+ ]
82
+ },
83
+ "observation.state": {
84
+ "type": "STATE",
85
+ "shape": [
86
+ 256
87
+ ]
88
+ },
89
+ "observation.task_info": {
90
+ "type": "STATE",
91
+ "shape": [
92
+ 46
93
+ ]
94
+ }
95
+ },
96
+ "output_features": {
97
+ "action": {
98
+ "type": "ACTION",
99
+ "shape": [
100
+ 23
101
+ ]
102
+ }
103
+ },
104
+ "device": "cuda",
105
+ "use_amp": false,
106
+ "push_to_hub": true,
107
+ "repo_id": "ywu67/pi05-task0000",
108
+ "private": null,
109
+ "tags": null,
110
+ "license": null,
111
+ "pretrained_path": null,
112
+ "paligemma_variant": "gemma_2b",
113
+ "action_expert_variant": "gemma_300m",
114
+ "dtype": "float32",
115
+ "chunk_size": 50,
116
+ "n_action_steps": 50,
117
+ "max_state_dim": 256,
118
+ "max_action_dim": 32,
119
+ "num_inference_steps": 10,
120
+ "time_sampling_beta_alpha": 1.5,
121
+ "time_sampling_beta_beta": 1.0,
122
+ "time_sampling_scale": 0.999,
123
+ "time_sampling_offset": 0.001,
124
+ "min_period": 0.004,
125
+ "max_period": 4.0,
126
+ "image_resolution": [
127
+ 224,
128
+ 224
129
+ ],
130
+ "empty_cameras": 0,
131
+ "tokenizer_max_length": 200,
132
+ "normalization_mapping": {
133
+ "VISUAL": "IDENTITY",
134
+ "STATE": "QUANTILES",
135
+ "ACTION": "QUANTILES"
136
+ },
137
+ "gradient_checkpointing": false,
138
+ "compile_model": false,
139
+ "compile_mode": "max-autotune",
140
+ "optimizer_lr": 2.5e-05,
141
+ "optimizer_betas": [
142
+ 0.9,
143
+ 0.95
144
+ ],
145
+ "optimizer_eps": 1e-08,
146
+ "optimizer_weight_decay": 0.01,
147
+ "optimizer_grad_clip_norm": 1.0,
148
+ "scheduler_warmup_steps": 1000,
149
+ "scheduler_decay_steps": 30000,
150
+ "scheduler_decay_lr": 2.5e-06
151
+ }
5000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9618a4a54eb8e2a847ba4a7218f63e793e59a35f37f273f8efe987c9c5bff420
3
+ size 14467170760
5000/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
+ 23
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": "cpu",
28
+ "float_dtype": null
29
+ }
30
+ }
31
+ ]
32
+ }
5000/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:089d5ed6f26ba87471187b591e4b793af3c976b96dabb5912b5ee36591a2770f
3
+ size 10924
5000/policy_preprocessor.json ADDED
@@ -0,0 +1,147 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.images.rgb.left_wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 480,
23
+ 480,
24
+ 3
25
+ ]
26
+ },
27
+ "observation.images.rgb.right_wrist": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 480,
31
+ 480,
32
+ 3
33
+ ]
34
+ },
35
+ "observation.images.rgb.head": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 720,
39
+ 720,
40
+ 3
41
+ ]
42
+ },
43
+ "observation.images.depth.left_wrist": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 480,
47
+ 480,
48
+ 3
49
+ ]
50
+ },
51
+ "observation.images.depth.right_wrist": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 480,
55
+ 480,
56
+ 3
57
+ ]
58
+ },
59
+ "observation.images.depth.head": {
60
+ "type": "VISUAL",
61
+ "shape": [
62
+ 720,
63
+ 720,
64
+ 3
65
+ ]
66
+ },
67
+ "observation.images.seg_instance_id.left_wrist": {
68
+ "type": "VISUAL",
69
+ "shape": [
70
+ 480,
71
+ 480,
72
+ 3
73
+ ]
74
+ },
75
+ "observation.images.seg_instance_id.right_wrist": {
76
+ "type": "VISUAL",
77
+ "shape": [
78
+ 480,
79
+ 480,
80
+ 3
81
+ ]
82
+ },
83
+ "observation.images.seg_instance_id.head": {
84
+ "type": "VISUAL",
85
+ "shape": [
86
+ 720,
87
+ 720,
88
+ 3
89
+ ]
90
+ },
91
+ "observation.cam_rel_poses": {
92
+ "type": "STATE",
93
+ "shape": [
94
+ 21
95
+ ]
96
+ },
97
+ "observation.state": {
98
+ "type": "STATE",
99
+ "shape": [
100
+ 256
101
+ ]
102
+ },
103
+ "observation.task_info": {
104
+ "type": "STATE",
105
+ "shape": [
106
+ 46
107
+ ]
108
+ },
109
+ "action": {
110
+ "type": "ACTION",
111
+ "shape": [
112
+ 23
113
+ ]
114
+ }
115
+ },
116
+ "norm_map": {
117
+ "VISUAL": "IDENTITY",
118
+ "STATE": "QUANTILES",
119
+ "ACTION": "QUANTILES"
120
+ }
121
+ },
122
+ "state_file": "policy_preprocessor_step_2_normalizer_processor.safetensors"
123
+ },
124
+ {
125
+ "registry_name": "pi05_prepare_state_tokenizer_processor_step",
126
+ "config": {}
127
+ },
128
+ {
129
+ "registry_name": "tokenizer_processor",
130
+ "config": {
131
+ "max_length": 200,
132
+ "task_key": "task",
133
+ "padding_side": "right",
134
+ "padding": "max_length",
135
+ "truncation": true,
136
+ "tokenizer_name": "google/paligemma-3b-pt-224"
137
+ }
138
+ },
139
+ {
140
+ "registry_name": "device_processor",
141
+ "config": {
142
+ "device": "cuda",
143
+ "float_dtype": null
144
+ }
145
+ }
146
+ ]
147
+ }
5000/policy_preprocessor_step_2_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:089d5ed6f26ba87471187b591e4b793af3c976b96dabb5912b5ee36591a2770f
3
+ size 10924
5000/train_config.json ADDED
@@ -0,0 +1,280 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "ywu67/2025-challenge-demos-task0000",
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": "pyav",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "pi05",
85
+ "n_obs_steps": 1,
86
+ "input_features": {
87
+ "observation.images.rgb.left_wrist": {
88
+ "type": "VISUAL",
89
+ "shape": [
90
+ 480,
91
+ 480,
92
+ 3
93
+ ]
94
+ },
95
+ "observation.images.rgb.right_wrist": {
96
+ "type": "VISUAL",
97
+ "shape": [
98
+ 480,
99
+ 480,
100
+ 3
101
+ ]
102
+ },
103
+ "observation.images.rgb.head": {
104
+ "type": "VISUAL",
105
+ "shape": [
106
+ 720,
107
+ 720,
108
+ 3
109
+ ]
110
+ },
111
+ "observation.images.depth.left_wrist": {
112
+ "type": "VISUAL",
113
+ "shape": [
114
+ 480,
115
+ 480,
116
+ 3
117
+ ]
118
+ },
119
+ "observation.images.depth.right_wrist": {
120
+ "type": "VISUAL",
121
+ "shape": [
122
+ 480,
123
+ 480,
124
+ 3
125
+ ]
126
+ },
127
+ "observation.images.depth.head": {
128
+ "type": "VISUAL",
129
+ "shape": [
130
+ 720,
131
+ 720,
132
+ 3
133
+ ]
134
+ },
135
+ "observation.images.seg_instance_id.left_wrist": {
136
+ "type": "VISUAL",
137
+ "shape": [
138
+ 480,
139
+ 480,
140
+ 3
141
+ ]
142
+ },
143
+ "observation.images.seg_instance_id.right_wrist": {
144
+ "type": "VISUAL",
145
+ "shape": [
146
+ 480,
147
+ 480,
148
+ 3
149
+ ]
150
+ },
151
+ "observation.images.seg_instance_id.head": {
152
+ "type": "VISUAL",
153
+ "shape": [
154
+ 720,
155
+ 720,
156
+ 3
157
+ ]
158
+ },
159
+ "observation.cam_rel_poses": {
160
+ "type": "STATE",
161
+ "shape": [
162
+ 21
163
+ ]
164
+ },
165
+ "observation.state": {
166
+ "type": "STATE",
167
+ "shape": [
168
+ 256
169
+ ]
170
+ },
171
+ "observation.task_info": {
172
+ "type": "STATE",
173
+ "shape": [
174
+ 46
175
+ ]
176
+ }
177
+ },
178
+ "output_features": {
179
+ "action": {
180
+ "type": "ACTION",
181
+ "shape": [
182
+ 23
183
+ ]
184
+ }
185
+ },
186
+ "device": "cuda",
187
+ "use_amp": false,
188
+ "push_to_hub": true,
189
+ "repo_id": "ywu67/pi05-task0000",
190
+ "private": null,
191
+ "tags": null,
192
+ "license": null,
193
+ "pretrained_path": null,
194
+ "paligemma_variant": "gemma_2b",
195
+ "action_expert_variant": "gemma_300m",
196
+ "dtype": "float32",
197
+ "chunk_size": 50,
198
+ "n_action_steps": 50,
199
+ "max_state_dim": 256,
200
+ "max_action_dim": 32,
201
+ "num_inference_steps": 10,
202
+ "time_sampling_beta_alpha": 1.5,
203
+ "time_sampling_beta_beta": 1.0,
204
+ "time_sampling_scale": 0.999,
205
+ "time_sampling_offset": 0.001,
206
+ "min_period": 0.004,
207
+ "max_period": 4.0,
208
+ "image_resolution": [
209
+ 224,
210
+ 224
211
+ ],
212
+ "empty_cameras": 0,
213
+ "tokenizer_max_length": 200,
214
+ "normalization_mapping": {
215
+ "VISUAL": "IDENTITY",
216
+ "STATE": "QUANTILES",
217
+ "ACTION": "QUANTILES"
218
+ },
219
+ "gradient_checkpointing": false,
220
+ "compile_model": false,
221
+ "compile_mode": "max-autotune",
222
+ "optimizer_lr": 2.5e-05,
223
+ "optimizer_betas": [
224
+ 0.9,
225
+ 0.95
226
+ ],
227
+ "optimizer_eps": 1e-08,
228
+ "optimizer_weight_decay": 0.01,
229
+ "optimizer_grad_clip_norm": 1.0,
230
+ "scheduler_warmup_steps": 1000,
231
+ "scheduler_decay_steps": 30000,
232
+ "scheduler_decay_lr": 2.5e-06
233
+ },
234
+ "output_dir": "/data/output/pi05_task0000",
235
+ "job_name": "pi05_task0000_100k",
236
+ "resume": false,
237
+ "seed": 1000,
238
+ "num_workers": 16,
239
+ "batch_size": 3,
240
+ "steps": 100000,
241
+ "eval_freq": 20000,
242
+ "log_freq": 200,
243
+ "save_checkpoint": true,
244
+ "save_freq": 5000,
245
+ "use_policy_training_preset": true,
246
+ "optimizer": {
247
+ "type": "adamw",
248
+ "lr": 2.5e-05,
249
+ "weight_decay": 0.01,
250
+ "grad_clip_norm": 1.0,
251
+ "betas": [
252
+ 0.9,
253
+ 0.95
254
+ ],
255
+ "eps": 1e-08
256
+ },
257
+ "scheduler": {
258
+ "type": "cosine_decay_with_warmup",
259
+ "num_warmup_steps": 1000,
260
+ "num_decay_steps": 30000,
261
+ "peak_lr": 2.5e-05,
262
+ "decay_lr": 2.5e-06
263
+ },
264
+ "eval": {
265
+ "n_episodes": 50,
266
+ "batch_size": 50,
267
+ "use_async_envs": false
268
+ },
269
+ "wandb": {
270
+ "enable": true,
271
+ "disable_artifact": true,
272
+ "project": "lerobot",
273
+ "entity": null,
274
+ "notes": null,
275
+ "run_id": "b49a27ot",
276
+ "mode": null
277
+ },
278
+ "checkpoint_path": null,
279
+ "rename_map": {}
280
+ }