GGUF My Repo

Takes any Hugging Face model repo and converts it to GGUF format on the spot, so you can run it locally with llama.cpp or any GGML-compatible runtime without touching a terminal. The gap it fills is real: most interesting fine-tunes ship in safetensors or pytorch bin, and the conversion pipeline has traditionally required a working Python environment, the right dependencies, and enough free disk space to not swear out loud. This Space wraps all of that into a single form. The output goes straight back to HF as a new repo under your account, quantized at whatever level you pick. No reservation on the concept — it does exactly one thing and does it well. The only friction is queue times when the Space is busy. -> Best for: indie hacker or AI engineer running local inference on fine-tuned models