#inference-engine

llama.cpp: The definitive LLM inference engine (MIT, 75k stars)

The definitive LLM inference engine — running quantized models on CPU and GPU with state-of-the-art performance via GGUF format.