Hugging Face has added a hardware selector to its Models page, allowing users to filter models compatible with their specific setup, such as an RTX 3060, an AMD processor, or an Apple Silicon M2 chip, based on account settings. The filter uses the size of available files, particularly quantized versions in GGUF format, and the RAM or VRAM of the declared setup, with the hardware database built from the 300,000 members who have agreed to share their hardware specs. Users can then obtain commands via the "Use this model" button and run models with tools like llama.cpp, Ollama, or LM Studio. For users with exotic chips, NPU accelerators, or newly released cards, they may need to wait or request additions via the forum.