songhan / DSD

DSD model zoo. Better accuracy models from DSD training on Imagenet with same model architecture

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

DSD Model Zoo

This repo contains pre-trained models by Dense-Sparse-Dense(DSD) training on Imagenet.

Compared to conventional training method, dense→sparse→dense (DSD) training yielded higher accuracy with same model architecture.

Sparsity is a powerful form of regularization. Our intuition is that, once the network arrives at a local minimum given the sparsity constraint, relaxing the constraint gives the network more freedom to escape the saddle point and arrive at a higher-accuracy local minimum.

Feel free to use the better-accuracy DSD models to help your research. If you find DSD traing useful, please cite the following paper:

@article{han2016_DSD,
  title={DSD: Dense-Sparse-Dense Training for Deep Neural Networks},
  author={Song Han, Jeff Pool, Sharan Narang, Huizi Mao, Enhao Gong, Shijian Tang, Erich Elsen, Peter Vajda, Manohar Paluri, John Tran, Bryan Catanzaro, William J. Dally},
  journal={International Conference on Learning Representations (ICLR)},
  year={2017}
}

Download:

AlexNet_DSD

VGG16_DSD

GoogleNet_DSD

SqueezeNet_DSD

ResNet18_DSD

ResNet50_DSD

Single-crop (224x224) validation error rate:

Baseline       Top-1 error Top-5 error DSD       Top-1 error Top-5 error
AlexNet   42.78%     19.73%     AlexNet_DSD   41.48%     18.71%
VGG16         31.50%      11.32%      VGG16_DSD 27.19% 8.67%
GoogleNet 31.14%     10.96%     GoogleNet_DSD 30.02% 10.34%
SqueezeNet   42.56%       19.52%       SqueezeNet_DSD 38.24% 16.53%
ResNet18     30.43%     10.76%   ResNet18_DSD 29.17% 10.13%
ResNet50       24.01%     7.02%       ResNet50_DSD 22.89%     6.47%

The beseline of AlexNet, VGG16, GoogleNet, SqueezeNet are from Caffe Model Zoo. The baseline of ResNet18, ResNet50 are from fb.resnet.torch commit 500b698.

About

DSD model zoo. Better accuracy models from DSD training on Imagenet with same model architecture

License:BSD 2-Clause "Simplified" License


Languages

Language:CSS 93.9%Language:JavaScript 5.1%Language:Shell 1.1%