Benchmarks of vLLM, InfiniGen, and H2O identify conditions under which each KV cache strategy delivers the best trade-off between memory consumption and inference performance.
The llama 3 herd of models,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Comparative Characterization of KV Cache Management Strategies for LLM Inference
Benchmarks of vLLM, InfiniGen, and H2O identify conditions under which each KV cache strategy delivers the best trade-off between memory consumption and inference performance.