Skip to content

PyTorch implementation of Language model compression with weighted low-rank factorization

Notifications You must be signed in to change notification settings

RahulSChand/Weighted-low-rank-factorization-Pytorch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Weighted-low-rank-factorization-Pytorch

PyTorch implementation of the paper Language model compression with weighted low-rank factorization (https://arxiv.org/abs/2207.00112 ICLR 2022)

Open source implementation of FWSVD(Factorized weighted Singular Value decompisition) training method introduced in https://arxiv.org/abs/2207.00112 using pytorch, accelerate & HuggingFace

Run train_glue_fwsvd.sh to run a pre-trained BERT 12 layer model on GLUE tasks with FWSVD. Use TASK_NAME flag to change the task

About

PyTorch implementation of Language model compression with weighted low-rank factorization

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published