Hugging Face's logo

mshukor
/
pppi05gshv25678states

Safetensors
Model card Files
xet
Community
mshukor commited on
Commit
c122a42
·
verified ·
1 Parent(s): 6f2219f

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. config.json +83 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +53 -0
  4. policy_preprocessor.json +78 -0
  5. train_config.json +234 -0
config.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "pi05",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 8
9
+ ]
10
+ },
11
+ "observation.images.front": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 480,
16
+ 640
17
+ ]
18
+ },
19
+ "observation.images.wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 480,
24
+ 640
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 7
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "push_to_hub": false,
39
+ "repo_id": null,
40
+ "private": null,
41
+ "tags": null,
42
+ "license": null,
43
+ "pretrained_path": "lerobot/pi05_base",
44
+ "paligemma_variant": "gemma_2b",
45
+ "action_expert_variant": "gemma_300m",
46
+ "dtype": "bfloat16",
47
+ "chunk_size": 50,
48
+ "n_action_steps": 50,
49
+ "max_state_dim": 32,
50
+ "max_action_dim": 32,
51
+ "num_inference_steps": 10,
52
+ "time_sampling_beta_alpha": 1.5,
53
+ "time_sampling_beta_beta": 1.0,
54
+ "time_sampling_scale": 0.999,
55
+ "time_sampling_offset": 0.001,
56
+ "min_period": 0.004,
57
+ "max_period": 4.0,
58
+ "image_resolution": [
59
+ 224,
60
+ 224
61
+ ],
62
+ "empty_cameras": 0,
63
+ "tokenizer_max_length": 200,
64
+ "normalization_mapping": {
65
+ "VISUAL": "IDENTITY",
66
+ "STATE": "QUANTILES",
67
+ "ACTION": "QUANTILES"
68
+ },
69
+ "gradient_checkpointing": false,
70
+ "compile_model": true,
71
+ "compile_mode": "max-autotune",
72
+ "optimizer_lr": 2.5e-05,
73
+ "optimizer_betas": [
74
+ 0.9,
75
+ 0.95
76
+ ],
77
+ "optimizer_eps": 1e-08,
78
+ "optimizer_weight_decay": 0.01,
79
+ "optimizer_grad_clip_norm": 1.0,
80
+ "scheduler_warmup_steps": 1000,
81
+ "scheduler_decay_steps": 30000,
82
+ "scheduler_decay_lr": 2.5e-06
83
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5609d4f9810dbc7ec6506c6895e929b044beca4c5b3984f7b3c469f8775e97e9
3
+ size 7473096344
policy_postprocessor.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "unnormalizer_processor",
6
+ "config": {
7
+ "eps": 1e-08,
8
+ "features": {
9
+ "observation.state": {
10
+ "type": "STATE",
11
+ "shape": [
12
+ 8
13
+ ]
14
+ },
15
+ "observation.images.front": {
16
+ "type": "VISUAL",
17
+ "shape": [
18
+ 3,
19
+ 480,
20
+ 640
21
+ ]
22
+ },
23
+ "observation.images.wrist": {
24
+ "type": "VISUAL",
25
+ "shape": [
26
+ 3,
27
+ 480,
28
+ 640
29
+ ]
30
+ },
31
+ "action": {
32
+ "type": "ACTION",
33
+ "shape": [
34
+ 7
35
+ ]
36
+ }
37
+ },
38
+ "norm_map": {
39
+ "VISUAL": "IDENTITY",
40
+ "STATE": "QUANTILES",
41
+ "ACTION": "QUANTILES"
42
+ }
43
+ }
44
+ },
45
+ {
46
+ "registry_name": "device_processor",
47
+ "config": {
48
+ "device": "cpu",
49
+ "float_dtype": null
50
+ }
51
+ }
52
+ ]
53
+ }
policy_preprocessor.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 8
23
+ ]
24
+ },
25
+ "observation.images.front": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 480,
30
+ 640
31
+ ]
32
+ },
33
+ "observation.images.wrist": {
34
+ "type": "VISUAL",
35
+ "shape": [
36
+ 3,
37
+ 480,
38
+ 640
39
+ ]
40
+ },
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 7
45
+ ]
46
+ }
47
+ },
48
+ "norm_map": {
49
+ "VISUAL": "IDENTITY",
50
+ "STATE": "QUANTILES",
51
+ "ACTION": "QUANTILES"
52
+ }
53
+ }
54
+ },
55
+ {
56
+ "registry_name": "pi05_prepare_state_tokenizer_processor_step",
57
+ "config": {}
58
+ },
59
+ {
60
+ "registry_name": "tokenizer_processor",
61
+ "config": {
62
+ "max_length": 200,
63
+ "task_key": "task",
64
+ "padding_side": "right",
65
+ "padding": "max_length",
66
+ "truncation": true,
67
+ "tokenizer_name": "google/paligemma-3b-pt-224"
68
+ }
69
+ },
70
+ {
71
+ "registry_name": "device_processor",
72
+ "config": {
73
+ "device": "cuda",
74
+ "float_dtype": null
75
+ }
76
+ }
77
+ ]
78
+ }
train_config.json ADDED
@@ -0,0 +1,234 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "mshukor/piper_pickplace_greenshampoo_v2,mickael-chen/piper_pickplace_greenshampoo_v5,mickael-chen/piper_pickplace_greenshampoo_v6,mickael-chen/piper_pickplace_greenshampoo_v7,mickael-chen/piper_pickplace_greenshampoo_v8",
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
+ }
62
+ },
63
+ "revision": null,
64
+ "use_imagenet_stats": false,
65
+ "video_backend": "torchcodec",
66
+ "streaming": false,
67
+ "sampling_weights": null,
68
+ "feature_keys_mapping": null,
69
+ "max_action_dim": null,
70
+ "max_state_dim": null,
71
+ "max_image_dim": null,
72
+ "max_num_images": null,
73
+ "train_on_all_features": false,
74
+ "min_fps": 1,
75
+ "max_fps": 100,
76
+ "ignore_keys": null
77
+ },
78
+ "env": null,
79
+ "policy": {
80
+ "type": "pi05",
81
+ "n_obs_steps": 1,
82
+ "input_features": {
83
+ "observation.state": {
84
+ "type": "STATE",
85
+ "shape": [
86
+ 8
87
+ ]
88
+ },
89
+ "observation.images.front": {
90
+ "type": "VISUAL",
91
+ "shape": [
92
+ 3,
93
+ 480,
94
+ 640
95
+ ]
96
+ },
97
+ "observation.images.wrist": {
98
+ "type": "VISUAL",
99
+ "shape": [
100
+ 3,
101
+ 480,
102
+ 640
103
+ ]
104
+ }
105
+ },
106
+ "output_features": {
107
+ "action": {
108
+ "type": "ACTION",
109
+ "shape": [
110
+ 7
111
+ ]
112
+ }
113
+ },
114
+ "device": "cuda",
115
+ "use_amp": false,
116
+ "push_to_hub": false,
117
+ "repo_id": null,
118
+ "private": null,
119
+ "tags": null,
120
+ "license": null,
121
+ "pretrained_path": "lerobot/pi05_base",
122
+ "paligemma_variant": "gemma_2b",
123
+ "action_expert_variant": "gemma_300m",
124
+ "dtype": "bfloat16",
125
+ "chunk_size": 50,
126
+ "n_action_steps": 50,
127
+ "max_state_dim": 32,
128
+ "max_action_dim": 32,
129
+ "num_inference_steps": 10,
130
+ "time_sampling_beta_alpha": 1.5,
131
+ "time_sampling_beta_beta": 1.0,
132
+ "time_sampling_scale": 0.999,
133
+ "time_sampling_offset": 0.001,
134
+ "min_period": 0.004,
135
+ "max_period": 4.0,
136
+ "image_resolution": [
137
+ 224,
138
+ 224
139
+ ],
140
+ "empty_cameras": 0,
141
+ "tokenizer_max_length": 200,
142
+ "normalization_mapping": {
143
+ "VISUAL": "IDENTITY",
144
+ "STATE": "QUANTILES",
145
+ "ACTION": "QUANTILES"
146
+ },
147
+ "gradient_checkpointing": false,
148
+ "compile_model": true,
149
+ "compile_mode": "max-autotune",
150
+ "optimizer_lr": 2.5e-05,
151
+ "optimizer_betas": [
152
+ 0.9,
153
+ 0.95
154
+ ],
155
+ "optimizer_eps": 1e-08,
156
+ "optimizer_weight_decay": 0.01,
157
+ "optimizer_grad_clip_norm": 1.0,
158
+ "scheduler_warmup_steps": 1000,
159
+ "scheduler_decay_steps": 30000,
160
+ "scheduler_decay_lr": 2.5e-06
161
+ },
162
+ "output_dir": "/lustre/fswork/projects/rech/dyf/ugz83ue/logs/nexus/lerobot_piper_pickplace_greenshampoo_v25678_pi05_bs16_onlyposstates",
163
+ "job_name": "pi05",
164
+ "resume": false,
165
+ "seed": 1000,
166
+ "num_workers": 4,
167
+ "batch_size": 16,
168
+ "steps": 200000,
169
+ "eval_freq": 20000,
170
+ "log_freq": 200,
171
+ "save_checkpoint": true,
172
+ "save_freq": 25000,
173
+ "use_policy_training_preset": true,
174
+ "optimizer": {
175
+ "type": "adamw",
176
+ "lr": 2.5e-05,
177
+ "weight_decay": 0.01,
178
+ "grad_clip_norm": 1.0,
179
+ "betas": [
180
+ 0.9,
181
+ 0.95
182
+ ],
183
+ "eps": 1e-08
184
+ },
185
+ "scheduler": {
186
+ "type": "cosine_decay_with_warmup",
187
+ "num_warmup_steps": 1000,
188
+ "num_decay_steps": 30000,
189
+ "peak_lr": 2.5e-05,
190
+ "decay_lr": 2.5e-06
191
+ },
192
+ "eval": {
193
+ "n_episodes": 50,
194
+ "batch_size": 50,
195
+ "use_async_envs": false
196
+ },
197
+ "wandb": {
198
+ "enable": false,
199
+ "disable_artifact": false,
200
+ "project": "lerobot",
201
+ "entity": null,
202
+ "notes": null,
203
+ "run_id": null,
204
+ "mode": null
205
+ },
206
+ "fsdp": null,
207
+ "freeze": {
208
+ "enable": true,
209
+ "module_names": [
210
+ "paligemma_with_expert.paligemma.",
211
+ "paligemma_with_expert.paligemma.lm_head",
212
+ "paligemma_with_expert.gemma_expert.lm_head"
213
+ ],
214
+ "invert_selection": false
215
+ },
216
+ "rename_policy_keys": {
217
+ "rename_keys": {}
218
+ },
219
+ "feature_filters": {
220
+ "feature_filters": {
221
+ "observation.state": {
222
+ "exclude_names": [
223
+ "gripper_roll.eff",
224
+ "wrist_flex.eff",
225
+ "wrist_roll.eff",
226
+ "elbow_flex.eff",
227
+ "shoulder_pan.eff",
228
+ "shoulder_lift.eff"
229
+ ]
230
+ }
231
+ }
232
+ },
233
+ "eval_num_batches": -1
234
+ }