/ThiNet

caffe model of ICCV'17 paper - ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression https://arxiv.org/abs/1707.06342

ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression

Pretrained caffe model of ICCV'17 paper:

"ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression"

For more details, please see our project page: [ThiNet Project Page]

Code

Caffe Implementation of ThiNet

Models

224x224 center crop validation accuracy on ImageNet, tested on one M40 GPU with batch_size=32.

Model Top-1 Top-5 #Param. #FLOPs f./b. (ms)
ThiNet-GAP 67.34% 87.92% 8.32M 9.34B 71.73/145.51
ThiNet-Tiny 59.34% 81.97% 1.32M 2.01B 29.51/55.83

Note: These two models are trained with different image cropping method, see trainval.prototxt for more details.

Citation

If you find this work useful for your research, please cite:

@CONFERENCE{ThiNet_ICCV17,
  author={Jian-Hao Luo, Jianxin Wu, and Weiyao Lin},
  title={ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression},
  booktitle={ICCV},
  year = {2017},
  pages={5058-5066},
}

Contact

Feel free to contact me if you have any question (Jian-Hao Luo luojh@lamda.nju.edu.cn or jianhao920@gmail.com).