conftrace_

salient token identification

1 paper

Explore in graph

Co-occurring keywords · papers mentioning both

adaptive compression (1) attention score (1) large language model (1) normalized attention score (1) tokenwise quantization (1) kv cache quantization (1) flash attention (1) attention mechanism (1) model compression (1)

Papers

ZipCache: Accurate and Efficient KV Cache Quantization with Salient Token Identification NeurIPS 2024