zhangxin-xd / BPL

Official Pytorch implementation of Block-wise Partner Learning for Model Compression[Accepted by IEEE TNNLS 2023]

Block-wise Partner Learning for Model Compression

⭐ Official Pytorch implementation of Block-wise Partner Learning for Model Compression [Accepted by IEEE TNNLS 2023]

Getting Started

Download the repo:

git clone https://github.com/zhangxin-xd/BPL.git
cd BPL

Data Preparation

• Put NWPU-45 data to ~/data_nwpu.
• Put UCML-21 data to ~/data_ucml.

Pruning

CUDA_VISIBLE_DEVICES=0 python main.py --dataset nwpu-45 --data_path ~/data_nwpu --block_type shadow --pruning_rate 0.5

Citation

@ARTICLE{10237122,
  author={Zhang, Xin and Xie, Weiying and Li, Yunsong and Lei, Jie and Jiang, Kai and Fang, Leyuan and Du, Qian},
  journal={IEEE Transactions on Neural Networks and Learning Systems}, 
  title={Block-Wise Partner Learning for Model Compression}, 
  year={2023},
  volume={},
  number={},
  pages={1-14},
  doi={10.1109/TNNLS.2023.3306512}}