Hugging Face's logo

mshukor
/
lpipv2578colorsv01bs32s250g2svla

Safetensors
Model card Files
xet
Community
mshukor commited on
Commit
8946d42
·
verified ·
1 Parent(s): 8e7161d

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. config.json +89 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +53 -0
  4. policy_preprocessor.json +78 -0
  5. train_config.json +238 -0
config.json ADDED
@@ -0,0 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
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": "/lustre/fswork/projects/rech/dyf/ugz83ue/.cache/huggingface/hub/smolvla_base_fixed_migrated",
44
+ "chunk_size": 50,
45
+ "n_action_steps": 50,
46
+ "normalization_mapping": {
47
+ "VISUAL": "IDENTITY",
48
+ "STATE": "MEAN_STD",
49
+ "ACTION": "MEAN_STD"
50
+ },
51
+ "max_state_dim": 32,
52
+ "max_action_dim": 32,
53
+ "resize_imgs_with_padding": [
54
+ 512,
55
+ 512
56
+ ],
57
+ "empty_cameras": 0,
58
+ "adapt_to_pi_aloha": false,
59
+ "use_delta_joint_actions_aloha": false,
60
+ "tokenizer_max_length": 48,
61
+ "num_steps": 10,
62
+ "use_cache": true,
63
+ "freeze_vision_encoder": true,
64
+ "train_expert_only": true,
65
+ "train_state_proj": true,
66
+ "optimizer_lr": 0.0001,
67
+ "optimizer_betas": [
68
+ 0.9,
69
+ 0.95
70
+ ],
71
+ "optimizer_eps": 1e-08,
72
+ "optimizer_weight_decay": 1e-10,
73
+ "optimizer_grad_clip_norm": 10,
74
+ "scheduler_warmup_steps": 1000,
75
+ "scheduler_decay_steps": 30000,
76
+ "scheduler_decay_lr": 2.5e-06,
77
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
78
+ "load_vlm_weights": false,
79
+ "add_image_special_tokens": false,
80
+ "attention_mode": "cross_attn",
81
+ "prefix_length": -1,
82
+ "pad_language_to": "longest",
83
+ "num_expert_layers": -1,
84
+ "num_vlm_layers": 16,
85
+ "self_attn_every_n_layers": 2,
86
+ "expert_width_multiplier": 0.75,
87
+ "min_period": 0.004,
88
+ "max_period": 4.0
89
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e9d806ae6aaa2cf62667b680f914b65575a617a859f5ffcb26206830ac2dc90
3
+ size 1197789224
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": "MEAN_STD",
41
+ "ACTION": "MEAN_STD"
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": "smolvla_new_line_processor",
16
+ "config": {}
17
+ },
18
+ {
19
+ "registry_name": "tokenizer_processor",
20
+ "config": {
21
+ "max_length": 48,
22
+ "task_key": "task",
23
+ "padding_side": "right",
24
+ "padding": "max_length",
25
+ "truncation": true,
26
+ "tokenizer_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct"
27
+ }
28
+ },
29
+ {
30
+ "registry_name": "device_processor",
31
+ "config": {
32
+ "device": "cuda",
33
+ "float_dtype": null
34
+ }
35
+ },
36
+ {
37
+ "registry_name": "normalizer_processor",
38
+ "config": {
39
+ "eps": 1e-08,
40
+ "features": {
41
+ "observation.state": {
42
+ "type": "STATE",
43
+ "shape": [
44
+ 8
45
+ ]
46
+ },
47
+ "observation.images.front": {
48
+ "type": "VISUAL",
49
+ "shape": [
50
+ 3,
51
+ 480,
52
+ 640
53
+ ]
54
+ },
55
+ "observation.images.wrist": {
56
+ "type": "VISUAL",
57
+ "shape": [
58
+ 3,
59
+ 480,
60
+ 640
61
+ ]
62
+ },
63
+ "action": {
64
+ "type": "ACTION",
65
+ "shape": [
66
+ 7
67
+ ]
68
+ }
69
+ },
70
+ "norm_map": {
71
+ "VISUAL": "IDENTITY",
72
+ "STATE": "MEAN_STD",
73
+ "ACTION": "MEAN_STD"
74
+ }
75
+ }
76
+ }
77
+ ]
78
+ }
train_config.json ADDED
@@ -0,0 +1,238 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "mshukor/piper_pickplace_greenshampoo_v2,mickael-chen/piper_pickplace_greenshampoo_v5,mickael-chen/piper_pickplace_greenshampoo_v7,mickael-chen/piper_pickplace_greenshampoo_v8,mickael-chen/piper_pickplace_brown,mickael-chen/piper_pickplace_orange_v01,mickael-chen/piper_pickplace_green_v01,mickael-chen/piper_pickplace_white_v01",
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": "smolvla",
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": "/lustre/fswork/projects/rech/dyf/ugz83ue/.cache/huggingface/hub/smolvla_base_fixed_migrated",
122
+ "chunk_size": 50,
123
+ "n_action_steps": 50,
124
+ "normalization_mapping": {
125
+ "VISUAL": "IDENTITY",
126
+ "STATE": "MEAN_STD",
127
+ "ACTION": "MEAN_STD"
128
+ },
129
+ "max_state_dim": 32,
130
+ "max_action_dim": 32,
131
+ "resize_imgs_with_padding": [
132
+ 512,
133
+ 512
134
+ ],
135
+ "empty_cameras": 0,
136
+ "adapt_to_pi_aloha": false,
137
+ "use_delta_joint_actions_aloha": false,
138
+ "tokenizer_max_length": 48,
139
+ "num_steps": 10,
140
+ "use_cache": true,
141
+ "freeze_vision_encoder": true,
142
+ "train_expert_only": true,
143
+ "train_state_proj": true,
144
+ "optimizer_lr": 0.0001,
145
+ "optimizer_betas": [
146
+ 0.9,
147
+ 0.95
148
+ ],
149
+ "optimizer_eps": 1e-08,
150
+ "optimizer_weight_decay": 1e-10,
151
+ "optimizer_grad_clip_norm": 10,
152
+ "scheduler_warmup_steps": 1000,
153
+ "scheduler_decay_steps": 30000,
154
+ "scheduler_decay_lr": 2.5e-06,
155
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
156
+ "load_vlm_weights": false,
157
+ "add_image_special_tokens": false,
158
+ "attention_mode": "cross_attn",
159
+ "prefix_length": -1,
160
+ "pad_language_to": "longest",
161
+ "num_expert_layers": -1,
162
+ "num_vlm_layers": 16,
163
+ "self_attn_every_n_layers": 2,
164
+ "expert_width_multiplier": 0.75,
165
+ "min_period": 0.004,
166
+ "max_period": 4.0
167
+ },
168
+ "output_dir": "/lustre/fswork/projects/rech/dyf/ugz83ue/logs/nexus/lerobot_piper_pickplace_greenshampoo_v2578_multicolor_v01_smolvla_bs32_s300000_gpus2onlyposstates",
169
+ "job_name": "smolvla",
170
+ "resume": false,
171
+ "seed": 1000,
172
+ "num_workers": 4,
173
+ "batch_size": 32,
174
+ "steps": 300000,
175
+ "eval_freq": 20000,
176
+ "log_freq": 200,
177
+ "save_checkpoint": true,
178
+ "save_freq": 50000,
179
+ "use_policy_training_preset": true,
180
+ "optimizer": {
181
+ "type": "adamw",
182
+ "lr": 0.0001,
183
+ "weight_decay": 1e-10,
184
+ "grad_clip_norm": 10,
185
+ "betas": [
186
+ 0.9,
187
+ 0.95
188
+ ],
189
+ "eps": 1e-08
190
+ },
191
+ "scheduler": {
192
+ "type": "cosine_decay_with_warmup",
193
+ "num_warmup_steps": 1000,
194
+ "num_decay_steps": 30000,
195
+ "peak_lr": 0.0001,
196
+ "decay_lr": 2.5e-06
197
+ },
198
+ "eval": {
199
+ "n_episodes": 50,
200
+ "batch_size": 50,
201
+ "use_async_envs": false
202
+ },
203
+ "wandb": {
204
+ "enable": false,
205
+ "disable_artifact": false,
206
+ "project": "lerobot",
207
+ "entity": null,
208
+ "notes": null,
209
+ "run_id": null,
210
+ "mode": null
211
+ },
212
+ "fsdp": null,
213
+ "freeze": {
214
+ "enable": true,
215
+ "module_names": [
216
+ "vlm_with_expert.vlm."
217
+ ],
218
+ "invert_selection": false
219
+ },
220
+ "rename_policy_keys": {
221
+ "rename_keys": {}
222
+ },
223
+ "feature_filters": {
224
+ "feature_filters": {
225
+ "observation.state": {
226
+ "exclude_names": [
227
+ "gripper_roll.eff",
228
+ "wrist_flex.eff",
229
+ "wrist_roll.eff",
230
+ "elbow_flex.eff",
231
+ "shoulder_pan.eff",
232
+ "shoulder_lift.eff"
233
+ ]
234
+ }
235
+ }
236
+ },
237
+ "eval_num_batches": -1
238
+ }