AI.info

Feng Xue

Explore Feng Xue on AI.info.

• HeteroCache: A Dynamic Retrieval Approach to Heterogeneous KV Cache Compression for Long-Context LLM Inference