Break the Sequential Dependency of LLM Inference Using Lookahead Decoding

Re-assign community