EMS improves KV cache compression with a Global-Local importance score and a head-wise Evict-then-Merge strategy, outperforming prior methods on LongBench and Needle-in-a-Haystack.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance
EMS improves KV cache compression with a Global-Local importance score and a head-wise Evict-then-Merge strategy, outperforming prior methods on LongBench and Needle-in-a-Haystack.