Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads
语言:Jupyter Notebook · GitHub 星标:2,774 · Fork:206
话题:llm · llm-inference
在 GitHub 上查看原仓库