Hugging Face's logo

Beable
/
leroVLA

Robotics
LeRobot
Safetensors
smolvla
Model card Files
xet
Community
Beable commited on
Commit
804c4df
·
verified ·
1 Parent(s): db308c6

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +80 -0
  2. model.safetensors +3 -0
  3. train_config.json +194 -0
config.json ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "normalization_mapping": {
5
+ "VISUAL": "IDENTITY",
6
+ "STATE": "MEAN_STD",
7
+ "ACTION": "MEAN_STD"
8
+ },
9
+ "input_features": {
10
+ "observation.images.rgb": {
11
+ "type": "VISUAL",
12
+ "shape": [
13
+ 3,
14
+ 360,
15
+ 640
16
+ ]
17
+ },
18
+ "observation.state": {
19
+ "type": "STATE",
20
+ "shape": [
21
+ 6
22
+ ]
23
+ }
24
+ },
25
+ "output_features": {
26
+ "action": {
27
+ "type": "ACTION",
28
+ "shape": [
29
+ 6
30
+ ]
31
+ }
32
+ },
33
+ "device": "cuda",
34
+ "use_amp": false,
35
+ "push_to_hub": false,
36
+ "repo_id": null,
37
+ "private": null,
38
+ "tags": null,
39
+ "license": null,
40
+ "chunk_size": 50,
41
+ "n_action_steps": 50,
42
+ "max_state_dim": 32,
43
+ "max_action_dim": 32,
44
+ "resize_imgs_with_padding": [
45
+ 512,
46
+ 512
47
+ ],
48
+ "empty_cameras": 0,
49
+ "adapt_to_pi_aloha": false,
50
+ "use_delta_joint_actions_aloha": false,
51
+ "tokenizer_max_length": 48,
52
+ "num_steps": 10,
53
+ "use_cache": true,
54
+ "freeze_vision_encoder": true,
55
+ "train_expert_only": true,
56
+ "train_state_proj": true,
57
+ "optimizer_lr": 0.0001,
58
+ "optimizer_betas": [
59
+ 0.9,
60
+ 0.95
61
+ ],
62
+ "optimizer_eps": 1e-08,
63
+ "optimizer_weight_decay": 1e-10,
64
+ "optimizer_grad_clip_norm": 10,
65
+ "scheduler_warmup_steps": 1000,
66
+ "scheduler_decay_steps": 30000,
67
+ "scheduler_decay_lr": 2.5e-06,
68
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
69
+ "load_vlm_weights": false,
70
+ "add_image_special_tokens": false,
71
+ "attention_mode": "cross_attn",
72
+ "prefix_length": -1,
73
+ "pad_language_to": "longest",
74
+ "num_expert_layers": -1,
75
+ "num_vlm_layers": 16,
76
+ "self_attn_every_n_layers": 2,
77
+ "expert_width_multiplier": 0.75,
78
+ "min_period": 0.004,
79
+ "max_period": 4.0
80
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de1d0562a61e5a3310a1a486f38aea8707b3493667b62b5bd8afce52d32ca9c6
3
+ size 1197790016
train_config.json ADDED
@@ -0,0 +1,194 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Beable/lerobot-SOARM100-sim",
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": true,
65
+ "video_backend": "pyav",
66
+ "tolerance_s": 0.2,
67
+ "timestamps_check": "warn"
68
+ },
69
+ "env": null,
70
+ "policy": {
71
+ "type": "smolvla",
72
+ "n_obs_steps": 1,
73
+ "normalization_mapping": {
74
+ "VISUAL": "IDENTITY",
75
+ "STATE": "MEAN_STD",
76
+ "ACTION": "MEAN_STD"
77
+ },
78
+ "input_features": {
79
+ "observation.images.rgb": {
80
+ "type": "VISUAL",
81
+ "shape": [
82
+ 3,
83
+ 360,
84
+ 640
85
+ ]
86
+ },
87
+ "observation.state": {
88
+ "type": "STATE",
89
+ "shape": [
90
+ 6
91
+ ]
92
+ }
93
+ },
94
+ "output_features": {
95
+ "action": {
96
+ "type": "ACTION",
97
+ "shape": [
98
+ 6
99
+ ]
100
+ }
101
+ },
102
+ "device": "cuda",
103
+ "use_amp": false,
104
+ "push_to_hub": false,
105
+ "repo_id": null,
106
+ "private": null,
107
+ "tags": null,
108
+ "license": null,
109
+ "chunk_size": 50,
110
+ "n_action_steps": 50,
111
+ "max_state_dim": 32,
112
+ "max_action_dim": 32,
113
+ "resize_imgs_with_padding": [
114
+ 512,
115
+ 512
116
+ ],
117
+ "empty_cameras": 0,
118
+ "adapt_to_pi_aloha": false,
119
+ "use_delta_joint_actions_aloha": false,
120
+ "tokenizer_max_length": 48,
121
+ "num_steps": 10,
122
+ "use_cache": true,
123
+ "freeze_vision_encoder": true,
124
+ "train_expert_only": true,
125
+ "train_state_proj": true,
126
+ "optimizer_lr": 0.0001,
127
+ "optimizer_betas": [
128
+ 0.9,
129
+ 0.95
130
+ ],
131
+ "optimizer_eps": 1e-08,
132
+ "optimizer_weight_decay": 1e-10,
133
+ "optimizer_grad_clip_norm": 10,
134
+ "scheduler_warmup_steps": 1000,
135
+ "scheduler_decay_steps": 30000,
136
+ "scheduler_decay_lr": 2.5e-06,
137
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
138
+ "load_vlm_weights": false,
139
+ "add_image_special_tokens": false,
140
+ "attention_mode": "cross_attn",
141
+ "prefix_length": -1,
142
+ "pad_language_to": "longest",
143
+ "num_expert_layers": -1,
144
+ "num_vlm_layers": 16,
145
+ "self_attn_every_n_layers": 2,
146
+ "expert_width_multiplier": 0.75,
147
+ "min_period": 0.004,
148
+ "max_period": 4.0
149
+ },
150
+ "output_dir": "outputs/train/vla_ft",
151
+ "job_name": "smolvla_ft",
152
+ "resume": false,
153
+ "seed": 1000,
154
+ "num_workers": 4,
155
+ "batch_size": 32,
156
+ "steps": 20000,
157
+ "eval_freq": 20000,
158
+ "log_freq": 50,
159
+ "save_checkpoint": true,
160
+ "save_freq": 20000,
161
+ "use_policy_training_preset": true,
162
+ "optimizer": {
163
+ "type": "adamw",
164
+ "lr": 0.0001,
165
+ "weight_decay": 1e-10,
166
+ "grad_clip_norm": 10,
167
+ "betas": [
168
+ 0.9,
169
+ 0.95
170
+ ],
171
+ "eps": 1e-08
172
+ },
173
+ "scheduler": {
174
+ "type": "cosine_decay_with_warmup",
175
+ "num_warmup_steps": 1000,
176
+ "num_decay_steps": 30000,
177
+ "peak_lr": 0.0001,
178
+ "decay_lr": 2.5e-06
179
+ },
180
+ "eval": {
181
+ "n_episodes": 0,
182
+ "batch_size": 0,
183
+ "use_async_envs": false
184
+ },
185
+ "wandb": {
186
+ "enable": false,
187
+ "disable_artifact": false,
188
+ "project": "lerobot",
189
+ "entity": null,
190
+ "notes": null,
191
+ "run_id": null,
192
+ "mode": null
193
+ }
194
+ }