March 6, 2020

348 words 2 mins read

huawei-noah/AdderNet

huawei-noah/AdderNet

Code for paper " AdderNet: Do We Really Need Multiplications in Deep Learning?”

repo name huawei-noah/AdderNet
repo link https://github.com/huawei-noah/AdderNet
homepage
language Python
size (curr.) 1345 kB
stars (curr.) 288
created 2020-02-25
license BSD 3-Clause “New” or “Revised” License

AdderNet: Do We Really Need Multiplications in Deep Learning?

This code is an inference demo of CVPR 2020 paper AdderNet: Do We Really Need Multiplications in Deep Learning?

We present adder networks (AdderNets) to trade massive multiplications in deep neural networks, especially convolutional neural networks (CNNs), for much cheaper additions to reduce computation costs. In AdderNets, we take the L1-norm distance between filters and input feature as the output response. As a result, the proposed AdderNets can achieve 74.9% Top-1 accuracy 91.7% Top-5 accuracy using ResNet-50 on the ImageNet dataset without any multiplication in convolution layer.

Requirements

  • python 3
  • pytorch >= 1.1.0
  • torchvision

Preparation

You can follow pytorch/examples to prepare the ImageNet data.

The pretrained models are available in google drive or baidu cloud (access code:126b)

Usage

Run python test.py --data_dir 'path/to/imagenet_root/' to evaluate on ImageNet val set. You will achieve 74.9% Top accuracy and 91.7% Top-5 accuracy on the ImageNet dataset using ResNet-50.

Run python test.py --dataset cifar10 --model_dir models/ResNet20-AdderNet.pth --data_dir 'path/to/cifar10_root/' to evaluate on CIFAR-10. You will achieve 91.8% accuracy on the CIFAR-10 dataset using ResNet-20.

The inference of AdderNets is slow since the adder filters is implemented by torch.cdist. You can write cuda to achieve higher inference speed.

Citation

@article{AdderNet,
    title={AdderNet: Do We Really Need Multiplications in Deep Learning?},
    author={Chen, Hanting and Wang, Yunhe and Xu, Chunjing and Shi, Boxin and Xu, Chao and Tian, Qi and Xu, Chang},
    journal={CVPR},
    year={2020}
}

Contributing

We appreciate all contributions. If you are planning to contribute back bug-fixes, please do so without any further discussion.

If you plan to contribute new features, utility functions or extensions to the core, please first open an issue and discuss the feature with us. Sending a PR without discussion might end up resulting in a rejected PR, because we might be taking the core in a different direction than you might be aware of.

comments powered by Disqus