Instructions to use RonPlusSign/smolvla_image_transforms with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LeRobot
How to use RonPlusSign/smolvla_image_transforms with LeRobot:
# See https://github.com/huggingface/lerobot?tab=readme-ov-file#installation for more details git clone https://github.com/huggingface/lerobot.git cd lerobot pip install -e .[smolvla]
# Launch finetuning on your dataset python lerobot/scripts/train.py \ --policy.path=RonPlusSign/smolvla_image_transforms \ --dataset.repo_id=lerobot/svla_so101_pickplace \ --batch_size=64 \ --steps=20000 \ --output_dir=outputs/train/my_smolvla \ --job_name=my_smolvla_training \ --policy.device=cuda \ --wandb.enable=true
# Run the policy using the record function python -m lerobot.record \ --robot.type=so101_follower \ --robot.port=/dev/ttyACM0 \ # <- Use your port --robot.id=my_blue_follower_arm \ # <- Use your robot id --robot.cameras="{ front: {type: opencv, index_or_path: 8, width: 640, height: 480, fps: 30}}" \ # <- Use your cameras --dataset.single_task="Grasp a lego block and put it in the bin." \ # <- Use the same task description you used in your dataset recording --dataset.repo_id=HF_USER/dataset_name \ # <- This will be the dataset name on HF Hub --dataset.episode_time_s=50 \ --dataset.num_episodes=10 \ --policy.path=RonPlusSign/smolvla_image_transforms - Notebooks
- Google Colab
- Kaggle
Upload policy weights, train config and readme
Browse files- config.json +3 -3
- model.safetensors +1 -1
- train_config.json +7 -7
config.json
CHANGED
|
@@ -54,7 +54,7 @@
|
|
| 54 |
"private": null,
|
| 55 |
"tags": null,
|
| 56 |
"license": null,
|
| 57 |
-
"pretrained_path":
|
| 58 |
"chunk_size": 20,
|
| 59 |
"n_action_steps": 20,
|
| 60 |
"normalization_mapping": {
|
|
@@ -84,9 +84,9 @@
|
|
| 84 |
],
|
| 85 |
"optimizer_eps": 1e-08,
|
| 86 |
"optimizer_weight_decay": 1e-10,
|
| 87 |
-
"optimizer_grad_clip_norm": 10,
|
| 88 |
"scheduler_warmup_steps": 1000,
|
| 89 |
-
"scheduler_decay_steps":
|
| 90 |
"scheduler_decay_lr": 2.5e-06,
|
| 91 |
"vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
|
| 92 |
"load_vlm_weights": true,
|
|
|
|
| 54 |
"private": null,
|
| 55 |
"tags": null,
|
| 56 |
"license": null,
|
| 57 |
+
"pretrained_path": "outputs/train/smolvla_image_transforms/checkpoints/last/pretrained_model",
|
| 58 |
"chunk_size": 20,
|
| 59 |
"n_action_steps": 20,
|
| 60 |
"normalization_mapping": {
|
|
|
|
| 84 |
],
|
| 85 |
"optimizer_eps": 1e-08,
|
| 86 |
"optimizer_weight_decay": 1e-10,
|
| 87 |
+
"optimizer_grad_clip_norm": 10.0,
|
| 88 |
"scheduler_warmup_steps": 1000,
|
| 89 |
+
"scheduler_decay_steps": 130000,
|
| 90 |
"scheduler_decay_lr": 2.5e-06,
|
| 91 |
"vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
|
| 92 |
"load_vlm_weights": true,
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 906712520
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f27e25fe1c976e289e774dc5759168e3eea4aacc6cd5f5fbada294d52c4563c
|
| 3 |
size 906712520
|
train_config.json
CHANGED
|
@@ -136,7 +136,7 @@
|
|
| 136 |
"private": null,
|
| 137 |
"tags": null,
|
| 138 |
"license": null,
|
| 139 |
-
"pretrained_path":
|
| 140 |
"chunk_size": 20,
|
| 141 |
"n_action_steps": 20,
|
| 142 |
"normalization_mapping": {
|
|
@@ -166,9 +166,9 @@
|
|
| 166 |
],
|
| 167 |
"optimizer_eps": 1e-08,
|
| 168 |
"optimizer_weight_decay": 1e-10,
|
| 169 |
-
"optimizer_grad_clip_norm": 10,
|
| 170 |
"scheduler_warmup_steps": 1000,
|
| 171 |
-
"scheduler_decay_steps":
|
| 172 |
"scheduler_decay_lr": 2.5e-06,
|
| 173 |
"vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
|
| 174 |
"load_vlm_weights": true,
|
|
@@ -185,11 +185,11 @@
|
|
| 185 |
},
|
| 186 |
"output_dir": "outputs/train/smolvla_image_transforms",
|
| 187 |
"job_name": "smolvla",
|
| 188 |
-
"resume":
|
| 189 |
"seed": 1000,
|
| 190 |
"num_workers": 4,
|
| 191 |
"batch_size": 32,
|
| 192 |
-
"steps":
|
| 193 |
"eval_freq": 20000,
|
| 194 |
"log_freq": 200,
|
| 195 |
"save_checkpoint": true,
|
|
@@ -199,7 +199,7 @@
|
|
| 199 |
"type": "adamw",
|
| 200 |
"lr": 0.0001,
|
| 201 |
"weight_decay": 1e-10,
|
| 202 |
-
"grad_clip_norm": 10,
|
| 203 |
"betas": [
|
| 204 |
0.9,
|
| 205 |
0.95
|
|
@@ -227,6 +227,6 @@
|
|
| 227 |
"run_id": "lglx7jru",
|
| 228 |
"mode": null
|
| 229 |
},
|
| 230 |
-
"checkpoint_path":
|
| 231 |
"rename_map": {}
|
| 232 |
}
|
|
|
|
| 136 |
"private": null,
|
| 137 |
"tags": null,
|
| 138 |
"license": null,
|
| 139 |
+
"pretrained_path": "outputs/train/smolvla_image_transforms/checkpoints/last/pretrained_model",
|
| 140 |
"chunk_size": 20,
|
| 141 |
"n_action_steps": 20,
|
| 142 |
"normalization_mapping": {
|
|
|
|
| 166 |
],
|
| 167 |
"optimizer_eps": 1e-08,
|
| 168 |
"optimizer_weight_decay": 1e-10,
|
| 169 |
+
"optimizer_grad_clip_norm": 10.0,
|
| 170 |
"scheduler_warmup_steps": 1000,
|
| 171 |
+
"scheduler_decay_steps": 130000,
|
| 172 |
"scheduler_decay_lr": 2.5e-06,
|
| 173 |
"vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
|
| 174 |
"load_vlm_weights": true,
|
|
|
|
| 185 |
},
|
| 186 |
"output_dir": "outputs/train/smolvla_image_transforms",
|
| 187 |
"job_name": "smolvla",
|
| 188 |
+
"resume": true,
|
| 189 |
"seed": 1000,
|
| 190 |
"num_workers": 4,
|
| 191 |
"batch_size": 32,
|
| 192 |
+
"steps": 130005,
|
| 193 |
"eval_freq": 20000,
|
| 194 |
"log_freq": 200,
|
| 195 |
"save_checkpoint": true,
|
|
|
|
| 199 |
"type": "adamw",
|
| 200 |
"lr": 0.0001,
|
| 201 |
"weight_decay": 1e-10,
|
| 202 |
+
"grad_clip_norm": 10.0,
|
| 203 |
"betas": [
|
| 204 |
0.9,
|
| 205 |
0.95
|
|
|
|
| 227 |
"run_id": "lglx7jru",
|
| 228 |
"mode": null
|
| 229 |
},
|
| 230 |
+
"checkpoint_path": "outputs/train/smolvla_image_transforms/checkpoints/last",
|
| 231 |
"rename_map": {}
|
| 232 |
}
|