FineQuant: Unlocking Efficiency with Fine-Grained Weight-Only Quantization for LLMs.

Young Jin Kim Rawn Henry Raffy Fahim Hany Hassan Awadalla

Published in: CoRR (2023)

Keyphrases

fine grained
coarse grained
access control
tightly coupled
massively parallel
computational complexity
quantization error
metadata
web services
privacy preserving
data provenance
data lineage