DQ-BART

tags

Transformers, BART, NLP

paper

(Li et al. 2022)

Architecture

It is a distilled and quantized version of BART. It improves performance as well as the model size.

Bibliography

1. Zheng Li, Zijian Wang, Ming Tan, Ramesh Nallapati, Parminder Bhatia, Andrew Arnold, Bing Xiang, Dan Roth. March 21, 2022. "DQ-BART: Efficient Sequence-to-sequence Model via Joint Distillation and Quantization". arXiv. DOI.