ai-llm-inference

描述

LLM inference patterns — latency budgeting, caching, batching, quantization, and parallelism. Use when optimizing serving cost or tail latency.