#asynchronous inference

Hugging Face launches SmolVLA, an efficient and affordable vision-language-action model enabling real-time robotic control on low-cost hardware with open-source resources.