SpotServe: Serving Generative Large Language Models on Preemptible Instances.
Xupeng MiaoChunan ShiJiangfei DuanXiaoli XiDahua LinBin CuiZhihao JiaPublished in: ASPLOS (2) (2024)
Keyphrases
- language model
- language modeling
- probabilistic model
- document retrieval
- language modelling
- speech recognition
- n gram
- generative model
- retrieval model
- information retrieval
- test collection
- query expansion
- statistical language models
- language modeling framework
- context sensitive
- document ranking
- vector space model
- ad hoc information retrieval
- smoothing methods
- query terms
- word error rate
- query specific
- passage retrieval
- retrieval effectiveness
- language models for information retrieval
- expectation maximization
- spoken term detection
- language model for information retrieval