Enabling Lightweight Fine-tuning for Pre-trained Language Model Compression based on Matrix Product Operators.
Peiyu LiuZe-Feng GaoWayne Xin ZhaoZhi-Yuan XieZhong-Yi LuJi-Rong WenPublished in: CoRR (2021)
Keyphrases
- lightweight
- language model
- fine tuning
- pre trained
- language modeling
- training data
- probabilistic model
- n gram
- document retrieval
- speech recognition
- information retrieval
- training examples
- retrieval model
- context sensitive
- mixture model
- control signals
- query expansion
- test collection
- neural network
- relevance model
- learning algorithm
- translation model
- information extraction
- small number
- bayesian networks
- smoothing methods