Login / Signup

LocMoE+: Enhanced Router with Token Feature Awareness for Efficient LLM Pre-Training.

Jing LiZhijie SunDachao LinXuan HeYi LinBinfan ZhengLi ZengRongqian ZhaoXin Chen
Published in: CoRR (2024)
Keyphrases
  • feature vectors
  • training set
  • cost effective
  • end to end
  • machine learning
  • decision trees
  • case study
  • object recognition
  • artificial neural networks
  • training process