Verified record · from the answer index

serving-llms-vllm

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible end

skill:davila7/claude-code-templates/serving-llms-vllm · published by davila7 · source repository