Awesome

Discriminator-Cooperated Feature Map Distillation for GAN Compression

Overview

Prerequisites

• Linux
• Python 3
• CPU or NVIDIA GPU + CUDA CuDNN

Getting Started

Installation

• Clone this repo:

  git clone https://github.com/Anonymous-Con/DCD-official.git
  cd DCD
  

• Install dependencies.

  conda create -n DCD python=3.7
  conda activate DCD
  pip install torch==1.7.0 torchvision==0.8.0 torchaudio==0.7.0 
  pip install -r requirements.txt 
  

Data preparation

• edges2shoes

  • Download the dataset

  bash datasets/download_pix2pix_dataset.sh edges2shoes-r
  

  • Get the statistical information for the ground-truth images for your dataset to compute FID.

  bash datasets/download_real_stat.sh edges2shoes-r B
  

• horse2zebra

  • Download the dataset

  bash datasets/download_cyclegan_dataset.sh horse2zebra
  

  • Get the statistical information for the ground-truth images for your dataset to compute FID.

  bash datasets/download_real_stat.sh horse2zebra A
  bash datasets/download_real_stat.sh horse2zebra B
  

• summer2winter

  • Download the dataset

  bash datasets/download_cyclegan_dataset.sh summer2winter_yosemite
  

  • Get the statistical information for the ground-truth images for your dataset to compute FID from here

Pretrained Model

We provide a list of pre-trained models in link.

Training

• pretrained vgg16 we should prepare weights of a vgg16 to calculate the style loss

• train student model using OMGD Run the following script to train a resnet-style student on horse2zebra dataset, all scripts for cyclegan and pix2pix on horse2zebra,summer2winter,edges2shoes can be found in ./scripts

  bash scripts/cycle_gan/horse2zebra/distill.sh
  

Testing

• test student models, FID or mIoU will be calculated, take resnet-style generator on horse2zebra dataset as an example

  bash scripts/cycle_gan/horse2zebra/test.sh
  

Acknowledgements

Our code is developed based on GAN Compression