SmolVLA: Efficient Vision-Language-Action Model trained on Lerobot Community Data

In three linesHugging Face introduces SmolVLA, an efficient vision-language-action model trained on Lerobot community data. The model combines visual perception and language understanding to generate robotic actions, optimized for inference on resource-constrained devices.

Summary generated by Claude — human-verified