The open-source ecosystem that has developed around LLaMA includes quantization tools GPTQ and GGML, the fine-tuning framework LoRA, and inference engines llama.cpp and vLLM, enabling consumer-grade GPUs to run language models with considerable capability.