Login / Signup

Athena: Efficient Block-Wise Post-Training Quantization for Large Language Models Using Second-Order Matrix Derivative Information.

Yanshu WangWenyang HeTong Yang
Published in: CoRR (2024)
Keyphrases
  • language model
  • block wise
  • probabilistic model
  • higher order
  • speech recognition
  • information retrieval
  • language modeling
  • training set
  • co occurrence
  • query expansion
  • test collection