lerobot/smolvla_metaworld
Robotics Model
SmolVLA MetaWorld is a fine-tuned vision-language-action model for robotic manipulation, trained on the MetaWorld MT50 benchmark, based on the SmolVLA base model.
Catalog metadata
- Repository owner
- lerobot
- Source
- Hugging Face Models
- Task
- robotics
- Cameras
- 0
- License
- apache-2.0
Browse the catalog to compare robots, tasks, sensors, and licenses. Sign in to open protected download and model links.