Hugging Face's logo

eorikun
/
primitive_motion_model_phase

Robotics
LeRobot
Safetensors
smolvla
Model card Files
xet
 Community
eorikun commited on
Commit
abb86dc
·
verified ·
1 Parent(s): a7168d5

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +63 -0
  2. config.json +89 -0
  3. model.safetensors +3 -0
  4. train_config.json +225 -0
README.md ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: lerobot/smolvla_base
3
+ datasets: eorikun/primitive_motion_phase
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 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.wrist_cam": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 480,
10
+ 640
11
+ ]
12
+ },
13
+ "observation.state": {
14
+ "type": "STATE",
15
+ "shape": [
16
+ 6
17
+ ]
18
+ },
19
+ "observation.task_phase": {
20
+ "type": "STATE",
21
+ "shape": [
22
+ 1
23
+ ]
24
+ }
25
+ },
26
+ "output_features": {
27
+ "action": {
28
+ "type": "ACTION",
29
+ "shape": [
30
+ 6
31
+ ]
32
+ }
33
+ },
34
+ "device": "cuda",
35
+ "use_amp": true,
36
+ "use_peft": false,
37
+ "push_to_hub": true,
38
+ "repo_id": "eorikun/primitive_motion_model_phase",
39
+ "private": null,
40
+ "tags": null,
41
+ "license": null,
42
+ "pretrained_path": null,
43
+ "chunk_size": 10,
44
+ "n_action_steps": 10,
45
+ "normalization_mapping": {
46
+ "VISUAL": "IDENTITY",
47
+ "STATE": "MEAN_STD",
48
+ "ACTION": "MEAN_STD"
49
+ },
50
+ "max_state_dim": 32,
51
+ "max_action_dim": 32,
52
+ "resize_imgs_with_padding": [
53
+ 512,
54
+ 512
55
+ ],
56
+ "empty_cameras": 0,
57
+ "adapt_to_pi_aloha": false,
58
+ "use_delta_joint_actions_aloha": false,
59
+ "tokenizer_max_length": 48,
60
+ "num_steps": 10,
61
+ "use_cache": true,
62
+ "freeze_vision_encoder": false,
63
+ "train_expert_only": true,
64
+ "train_state_proj": true,
65
+ "optimizer_lr": 0.0001,
66
+ "optimizer_betas": [
67
+ 0.9,
68
+ 0.95
69
+ ],
70
+ "optimizer_eps": 1e-08,
71
+ "optimizer_weight_decay": 1e-10,
72
+ "optimizer_grad_clip_norm": 10,
73
+ "scheduler_warmup_steps": 1000,
74
+ "scheduler_decay_steps": 30000,
75
+ "scheduler_decay_lr": 2.5e-06,
76
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
77
+ "load_vlm_weights": false,
78
+ "add_image_special_tokens": false,
79
+ "attention_mode": "cross_attn",
80
+ "prefix_length": -1,
81
+ "pad_language_to": "longest",
82
+ "num_expert_layers": -1,
83
+ "num_vlm_layers": 16,
84
+ "self_attn_every_n_layers": 2,
85
+ "expert_width_multiplier": 0.75,
86
+ "min_period": 0.004,
87
+ "max_period": 4.0,
88
+ "rtc_config": null
89
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8cfa4a997306a33154a71666de41bbf8214847b145337cd8107fea096b9e465d
3
+ size 1197789224
train_config.json ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "eorikun/primitive_motion_phase",
4
+ "root": "/home/jetson/datasets/primitive_motion_phase",
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": "smolvla",
85
+ "n_obs_steps": 1,
86
+ "input_features": {
87
+ "observation.images.wrist_cam": {
88
+ "type": "VISUAL",
89
+ "shape": [
90
+ 3,
91
+ 480,
92
+ 640
93
+ ]
94
+ },
95
+ "observation.state": {
96
+ "type": "STATE",
97
+ "shape": [
98
+ 6
99
+ ]
100
+ },
101
+ "observation.task_phase": {
102
+ "type": "STATE",
103
+ "shape": [
104
+ 1
105
+ ]
106
+ }
107
+ },
108
+ "output_features": {
109
+ "action": {
110
+ "type": "ACTION",
111
+ "shape": [
112
+ 6
113
+ ]
114
+ }
115
+ },
116
+ "device": "cuda",
117
+ "use_amp": true,
118
+ "use_peft": false,
119
+ "push_to_hub": true,
120
+ "repo_id": "eorikun/primitive_motion_model_phase",
121
+ "private": null,
122
+ "tags": null,
123
+ "license": null,
124
+ "pretrained_path": null,
125
+ "chunk_size": 10,
126
+ "n_action_steps": 10,
127
+ "normalization_mapping": {
128
+ "VISUAL": "IDENTITY",
129
+ "STATE": "MEAN_STD",
130
+ "ACTION": "MEAN_STD"
131
+ },
132
+ "max_state_dim": 32,
133
+ "max_action_dim": 32,
134
+ "resize_imgs_with_padding": [
135
+ 512,
136
+ 512
137
+ ],
138
+ "empty_cameras": 0,
139
+ "adapt_to_pi_aloha": false,
140
+ "use_delta_joint_actions_aloha": false,
141
+ "tokenizer_max_length": 48,
142
+ "num_steps": 10,
143
+ "use_cache": true,
144
+ "freeze_vision_encoder": false,
145
+ "train_expert_only": true,
146
+ "train_state_proj": true,
147
+ "optimizer_lr": 0.0001,
148
+ "optimizer_betas": [
149
+ 0.9,
150
+ 0.95
151
+ ],
152
+ "optimizer_eps": 1e-08,
153
+ "optimizer_weight_decay": 1e-10,
154
+ "optimizer_grad_clip_norm": 10,
155
+ "scheduler_warmup_steps": 1000,
156
+ "scheduler_decay_steps": 30000,
157
+ "scheduler_decay_lr": 2.5e-06,
158
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
159
+ "load_vlm_weights": false,
160
+ "add_image_special_tokens": false,
161
+ "attention_mode": "cross_attn",
162
+ "prefix_length": -1,
163
+ "pad_language_to": "longest",
164
+ "num_expert_layers": -1,
165
+ "num_vlm_layers": 16,
166
+ "self_attn_every_n_layers": 2,
167
+ "expert_width_multiplier": 0.75,
168
+ "min_period": 0.004,
169
+ "max_period": 4.0,
170
+ "rtc_config": null
171
+ },
172
+ "output_dir": "outputs/primitive_motion_phase_30k",
173
+ "job_name": "smolvla",
174
+ "resume": false,
175
+ "seed": 1000,
176
+ "num_workers": 2,
177
+ "batch_size": 4,
178
+ "steps": 30000,
179
+ "eval_freq": 20000,
180
+ "log_freq": 50,
181
+ "tolerance_s": 0.0001,
182
+ "save_checkpoint": true,
183
+ "save_freq": 5000,
184
+ "use_policy_training_preset": true,
185
+ "optimizer": {
186
+ "type": "adamw",
187
+ "lr": 0.0001,
188
+ "weight_decay": 1e-10,
189
+ "grad_clip_norm": 10,
190
+ "betas": [
191
+ 0.9,
192
+ 0.95
193
+ ],
194
+ "eps": 1e-08
195
+ },
196
+ "scheduler": {
197
+ "type": "cosine_decay_with_warmup",
198
+ "num_warmup_steps": 1000,
199
+ "num_decay_steps": 30000,
200
+ "peak_lr": 0.0001,
201
+ "decay_lr": 2.5e-06
202
+ },
203
+ "eval": {
204
+ "n_episodes": 50,
205
+ "batch_size": 50,
206
+ "use_async_envs": false
207
+ },
208
+ "wandb": {
209
+ "enable": true,
210
+ "disable_artifact": false,
211
+ "project": "primitive_motion",
212
+ "entity": null,
213
+ "notes": null,
214
+ "run_id": "ve61o0e6",
215
+ "mode": null
216
+ },
217
+ "peft": null,
218
+ "use_rabc": false,
219
+ "rabc_progress_path": null,
220
+ "rabc_kappa": 0.01,
221
+ "rabc_epsilon": 1e-06,
222
+ "rabc_head_mode": "sparse",
223
+ "rename_map": {},
224
+ "checkpoint_path": null
225
+ }