snu-mllab/KVzip

[NeurIPS'25 Oral] Query-agnostic KV cache eviction: 3–4× reduction in memory and 2× decrease in latency (Qwen3/2.5, Gemma3, LLaMA3)

3 best points · 0 comments · 1 Hacker News post · 228 GitHub stars · View on GitHub

Hacker News discussions

• KVzip: Query-Agnostic KV Cache Compression with Context Reconstruction (3 points, 0 comments, by biztos, 2025-11-13)