🚨 SmolVLA is here — and it’s changing how we think about robotics AI. Hugging Face just released SmolVLA, a lightweight Vision-Language-Action model trained on community-shared datasets from their LeRobot platform. Despite being just 450M parameters, it outperforms larger models in both simulated and real-world robotics tasks — and it runs on a MacBook or single consumer GPU. 🤯 But it’s not just the model that’s impressive. It’s the movement behind it: -Built on open-source robotics datasets -Optimized for affordable hardware -Designed for real-world action, not just lab demos -Supports asynchronous inference, enabling faster reaction times in dynamic environments SmolVLA proves that powerful robotics AI doesn’t have to be heavy, expensive, or locked behind research walls. We're entering a new era where garage tinkerers, indie researchers, and global devs can build, test, and deploy intelligent robots — all thanks to community-powered ecosystems. 👀 The robotics revolution won’t be centralized. It’ll be open-source, async, and probably running on someone’s laptop. #AI #Robotics #HuggingFace #SmolVLA #OpenSource
Download the medial app to read full posts, comements and news.