Hugging Face's logo

robot-learning-group47
/
eval1_red_smolvla

Robotics
LeRobot
Safetensors
smolvla
Model card Files
xet
 Community
nkim7 commited on
Commit
a8f1eba
·
verified ·
1 Parent(s): bacb6a7

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +63 -0
  2. config.json +117 -0
  3. model.safetensors +3 -0
  4. train_config.json +257 -0
README.md ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: lerobot/smolvla_base
3
+ datasets: robot-learning-group47/eval1_red
4
+ library_name: lerobot
5
+ license: apache-2.0
6
+ model_name: smolvla
7
+ pipeline_tag: robotics
8
+ tags:
9
+ - robotics
10
+ - smolvla
11
+ - lerobot
12
+ ---
13
+
14
+ # Model Card for smolvla
15
+
16
+ <!-- Provide a quick summary of what the model is/does. -->
17
+
18
+
19
+ [SmolVLA](https://huggingface.co/papers/2506.01844) is a compact, efficient vision-language-action model that achieves competitive performance at reduced computational costs and can be deployed on consumer-grade hardware.
20
+
21
+
22
+ This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
23
+ See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
24
+
25
+ ---
26
+
27
+ ## How to Get Started with the Model
28
+
29
+ For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
30
+ Below is the short version on how to train and run inference/eval:
31
+
32
+ ### Train from scratch
33
+
34
+ ```bash
35
+ lerobot-train \
36
+ --dataset.repo_id=${HF_USER}/<dataset> \
37
+ --policy.type=act \
38
+ --output_dir=outputs/train/<desired_policy_repo_id> \
39
+ --job_name=lerobot_training \
40
+ --policy.device=cuda \
41
+ --policy.repo_id=${HF_USER}/<desired_policy_repo_id>
42
+ --wandb.enable=true
43
+ ```
44
+
45
+ _Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
46
+
47
+ ### Evaluate the policy/run inference
48
+
49
+ ```bash
50
+ lerobot-record \
51
+ --robot.type=so100_follower \
52
+ --dataset.repo_id=<hf_user>/eval_<dataset> \
53
+ --policy.path=<hf_user>/<desired_policy_repo_id> \
54
+ --episodes=10
55
+ ```
56
+
57
+ Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
58
+
59
+ ---
60
+
61
+ ## Model Details
62
+
63
+ - **License:** apache-2.0
config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 6
9
+ ]
10
+ },
11
+ "observation.images.camera1": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.camera2": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.camera3": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ },
35
+ "observation.images.empty_camera_0": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 3,
39
+ 480,
40
+ 640
41
+ ]
42
+ },
43
+ "observation.images.empty_camera_1": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 480,
48
+ 640
49
+ ]
50
+ }
51
+ },
52
+ "output_features": {
53
+ "action": {
54
+ "type": "ACTION",
55
+ "shape": [
56
+ 6
57
+ ]
58
+ }
59
+ },
60
+ "device": "cuda",
61
+ "use_amp": false,
62
+ "use_peft": false,
63
+ "push_to_hub": true,
64
+ "repo_id": "robot-learning-group47/eval1_red_smolvla",
65
+ "private": null,
66
+ "tags": null,
67
+ "license": null,
68
+ "pretrained_path": "lerobot/smolvla_base",
69
+ "chunk_size": 50,
70
+ "n_action_steps": 50,
71
+ "normalization_mapping": {
72
+ "VISUAL": "IDENTITY",
73
+ "STATE": "MEAN_STD",
74
+ "ACTION": "MEAN_STD"
75
+ },
76
+ "max_state_dim": 32,
77
+ "max_action_dim": 32,
78
+ "resize_imgs_with_padding": [
79
+ 512,
80
+ 512
81
+ ],
82
+ "empty_cameras": 2,
83
+ "adapt_to_pi_aloha": false,
84
+ "use_delta_joint_actions_aloha": false,
85
+ "tokenizer_max_length": 48,
86
+ "num_steps": 10,
87
+ "use_cache": true,
88
+ "freeze_vision_encoder": true,
89
+ "train_expert_only": true,
90
+ "train_state_proj": true,
91
+ "optimizer_lr": 0.0001,
92
+ "optimizer_betas": [
93
+ 0.9,
94
+ 0.95
95
+ ],
96
+ "optimizer_eps": 1e-08,
97
+ "optimizer_weight_decay": 1e-10,
98
+ "optimizer_grad_clip_norm": 10.0,
99
+ "scheduler_warmup_steps": 1000,
100
+ "scheduler_decay_steps": 30000,
101
+ "scheduler_decay_lr": 2.5e-06,
102
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
103
+ "load_vlm_weights": true,
104
+ "add_image_special_tokens": false,
105
+ "attention_mode": "cross_attn",
106
+ "prefix_length": 0,
107
+ "pad_language_to": "max_length",
108
+ "num_expert_layers": 0,
109
+ "num_vlm_layers": 16,
110
+ "self_attn_every_n_layers": 2,
111
+ "expert_width_multiplier": 0.75,
112
+ "min_period": 0.004,
113
+ "max_period": 4.0,
114
+ "rtc_config": null,
115
+ "compile_model": false,
116
+ "compile_mode": "max-autotune"
117
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:347a791fcb1289da5fbd88c0bcc8f033acd172405d03ac564b8496ca66f8093e
3
+ size 906712520
train_config.json ADDED
@@ -0,0 +1,257 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "robot-learning-group47/eval1_red",
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": "torchcodec",
80
+ "return_uint8": false,
81
+ "streaming": false
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "smolvla",
86
+ "n_obs_steps": 1,
87
+ "input_features": {
88
+ "observation.state": {
89
+ "type": "STATE",
90
+ "shape": [
91
+ 6
92
+ ]
93
+ },
94
+ "observation.images.camera1": {
95
+ "type": "VISUAL",
96
+ "shape": [
97
+ 3,
98
+ 256,
99
+ 256
100
+ ]
101
+ },
102
+ "observation.images.camera2": {
103
+ "type": "VISUAL",
104
+ "shape": [
105
+ 3,
106
+ 256,
107
+ 256
108
+ ]
109
+ },
110
+ "observation.images.camera3": {
111
+ "type": "VISUAL",
112
+ "shape": [
113
+ 3,
114
+ 256,
115
+ 256
116
+ ]
117
+ },
118
+ "observation.images.empty_camera_0": {
119
+ "type": "VISUAL",
120
+ "shape": [
121
+ 3,
122
+ 480,
123
+ 640
124
+ ]
125
+ },
126
+ "observation.images.empty_camera_1": {
127
+ "type": "VISUAL",
128
+ "shape": [
129
+ 3,
130
+ 480,
131
+ 640
132
+ ]
133
+ }
134
+ },
135
+ "output_features": {
136
+ "action": {
137
+ "type": "ACTION",
138
+ "shape": [
139
+ 6
140
+ ]
141
+ }
142
+ },
143
+ "device": "cuda",
144
+ "use_amp": false,
145
+ "use_peft": false,
146
+ "push_to_hub": true,
147
+ "repo_id": "robot-learning-group47/eval1_red_smolvla",
148
+ "private": null,
149
+ "tags": null,
150
+ "license": null,
151
+ "pretrained_path": "lerobot/smolvla_base",
152
+ "chunk_size": 50,
153
+ "n_action_steps": 50,
154
+ "normalization_mapping": {
155
+ "VISUAL": "IDENTITY",
156
+ "STATE": "MEAN_STD",
157
+ "ACTION": "MEAN_STD"
158
+ },
159
+ "max_state_dim": 32,
160
+ "max_action_dim": 32,
161
+ "resize_imgs_with_padding": [
162
+ 512,
163
+ 512
164
+ ],
165
+ "empty_cameras": 2,
166
+ "adapt_to_pi_aloha": false,
167
+ "use_delta_joint_actions_aloha": false,
168
+ "tokenizer_max_length": 48,
169
+ "num_steps": 10,
170
+ "use_cache": true,
171
+ "freeze_vision_encoder": true,
172
+ "train_expert_only": true,
173
+ "train_state_proj": true,
174
+ "optimizer_lr": 0.0001,
175
+ "optimizer_betas": [
176
+ 0.9,
177
+ 0.95
178
+ ],
179
+ "optimizer_eps": 1e-08,
180
+ "optimizer_weight_decay": 1e-10,
181
+ "optimizer_grad_clip_norm": 10.0,
182
+ "scheduler_warmup_steps": 1000,
183
+ "scheduler_decay_steps": 30000,
184
+ "scheduler_decay_lr": 2.5e-06,
185
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
186
+ "load_vlm_weights": true,
187
+ "add_image_special_tokens": false,
188
+ "attention_mode": "cross_attn",
189
+ "prefix_length": 0,
190
+ "pad_language_to": "max_length",
191
+ "num_expert_layers": 0,
192
+ "num_vlm_layers": 16,
193
+ "self_attn_every_n_layers": 2,
194
+ "expert_width_multiplier": 0.75,
195
+ "min_period": 0.004,
196
+ "max_period": 4.0,
197
+ "rtc_config": null,
198
+ "compile_model": false,
199
+ "compile_mode": "max-autotune"
200
+ },
201
+ "reward_model": null,
202
+ "output_dir": "outputs/train/eval1_red_smolvla",
203
+ "job_name": "eval1_red_smolvla",
204
+ "resume": false,
205
+ "seed": 1000,
206
+ "cudnn_deterministic": false,
207
+ "num_workers": 4,
208
+ "batch_size": 32,
209
+ "prefetch_factor": 4,
210
+ "persistent_workers": true,
211
+ "steps": 5000,
212
+ "eval_freq": 20000,
213
+ "log_freq": 200,
214
+ "tolerance_s": 0.0001,
215
+ "save_checkpoint": true,
216
+ "save_freq": 1000,
217
+ "use_policy_training_preset": true,
218
+ "optimizer": {
219
+ "type": "adamw",
220
+ "lr": 0.0001,
221
+ "weight_decay": 1e-10,
222
+ "grad_clip_norm": 10.0,
223
+ "betas": [
224
+ 0.9,
225
+ 0.95
226
+ ],
227
+ "eps": 1e-08
228
+ },
229
+ "scheduler": {
230
+ "type": "cosine_decay_with_warmup",
231
+ "num_warmup_steps": 1000,
232
+ "num_decay_steps": 30000,
233
+ "peak_lr": 0.0001,
234
+ "decay_lr": 2.5e-06
235
+ },
236
+ "eval": {
237
+ "n_episodes": 50,
238
+ "batch_size": 8,
239
+ "use_async_envs": true
240
+ },
241
+ "wandb": {
242
+ "enable": true,
243
+ "disable_artifact": false,
244
+ "project": "lerobot",
245
+ "entity": null,
246
+ "notes": null,
247
+ "run_id": "sr8lr9oe",
248
+ "mode": null,
249
+ "add_tags": true
250
+ },
251
+ "peft": null,
252
+ "sample_weighting": null,
253
+ "rename_map": {
254
+ "observation.images.front": "observation.images.camera1"
255
+ },
256
+ "checkpoint_path": null
257
+ }