How do I quantize a Mistral LLM for better inference speed on low-end GPUs

Question

May i know How do I quantize a Mistral LLM for better inference speed on low-end GPUs?