A voting-based KV cache eviction algorithm, a reconfigurable GEMV dataflow, and an element-serial non-linear scheduler are combined in the VEDA accelerator to speed up edge LLM generation.
Keyformer: Kv cache reduction through key tokens selection for efficient generative inference,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AR 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
VEDA: Efficient LLM Generation Through Voting-based KV Cache Eviction and Dataflow-flexible Accelerator
A voting-based KV cache eviction algorithm, a reconfigurable GEMV dataflow, and an element-serial non-linear scheduler are combined in the VEDA accelerator to speed up edge LLM generation.