Rethinking Memory and Communication Cost for Efficient Large Language Model Training.
Chan WuHanxiao ZhangLin JuJinjing HuangYoushao XiaoZhaoxin HuanSiyuan LiFanzhuang MengLei LiangXiaolu ZhangJun ZhouPublished in: CoRR (2023)
Keyphrases
- language model
- communication cost
- language modeling
- n gram
- document retrieval
- speech recognition
- query expansion
- reduce communication cost
- probabilistic model
- statistical language models
- sensor networks
- mixture model
- context sensitive
- retrieval model
- query terms
- information retrieval
- translation model
- language modelling
- test collection
- ad hoc information retrieval
- data streams
- pseudo relevance feedback
- language model for information retrieval
- data sets
- multi dimensional
- query specific
- decision trees