AI & ML interests
Efficient AI
Papers
CompactAttention: Accelerating Chunked Prefill with Block-Union KV Selection
RelayGen: Intra-Generation Model Switching for Efficient Reasoning
SNU-VLSI 's models
None public yet