Hugging Face has introduced SmolVLA, an open AI model for robotics that claims to outperform larger models in both virtual and real environments. With 450 million parameters, SmolVLA democratizes access to advanced vision-language-action models, supporting research in generalist robotics. It can run on consumer GPUs and affordable hardware, making it accessible for personal projects. The model features an asynchronous inference stack that enhances response times by processing actions separately from sensory input. This initiative is part of Hugging Face's broader strategy to build a low-cost robotics ecosystem, including acquisitions and new robotics systems.
It's becoming a little easier to build sophisticated robotics projects at home.
SmolVLA aims to democratize access to vision-language-action models and accelerate research toward generalist robotic agents.
SmolVLA is small enough to run on a single consumer GPU - or even a MacBook - and can be tested and deployed on affordable hardware.
Asynchronous inference stack allows the model to separate the processing of a robot's actions from the processing of what it sees and hears.
Collection
[
|
...
]