Fast, Cheap, and Accurate: Optimizing LLM Inference with vLLM and Quantization Similarity score = 0.76 More