📈 GH Trending Explorer 🎲 Surprise me
lyogavin / airllm

Python library that runs large language model inference layer by layer, so 70B-class models fit on a single 4GB GPU without quantization.

▲ ~2.2k/7d ★ 40k Jupyter Notebook 🔥 today