AI.info

Gang Lin

Explore Gang Lin on AI.info.

• LycheeDecode: Accelerating Long-Context LLM Inference via Hybrid-Head Sparse Decoding