News/CompressKV: Semantic-Retrieval-Guided KV-Cache C…
SatelightUnresolved

CompressKV: Semantic-Retrieval-Guided KV-Cache Compression for Resource-Efficient Long-Context LLM Inference

Read full release on Satelight
  • TUDa-HWAI
CompressKV: Semantic-Retrieval-Guided KV-Cache Compression for Resource-Efficient Long-Context LLM Inference