Skip to content
Enhanced SRGAN. Champion PIRM Challenge on Perceptual Super-Resolution
Python
Branch: master
Clone or download
Latest commit b980f5e Jun 2, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore update: gitignore May 12, 2019
LICENSE
README.md
config.py update: type-hinting Jun 2, 2019
dataloader.py add: initial files May 12, 2019
losses.py feat: impl several GAN losses Jun 2, 2019
main.py
metrics.py
model.py update: remove some codes, not used currently Jun 2, 2019
requirements.txt add: required pip packages May 12, 2019
tfutils.py feat: impl pixel-shuffle Jun 2, 2019
utils.py
vgg19.py add: impl VGG19 model w/ loader Jun 2, 2019

README.md

ESRGAN in tensorflow

Enhanced Super Resolution Generative Adversarial Network in tensorflow

This repo is based on pytorch impl original here

Work In Process :)

Total alerts Language grade: Python

Requirements

  • python 2.x / 3.x
  • tensorflow-gpu 1.x
  • opencv
  • glob
  • tqdm

Repo-Tree

│
├── output  (generated images)
│     ├── ...
│     └── xxx.png
├── tb_logs (tensorboard records)
│     ├── [unique id]
│     │     ├── *.ckpt
│     │     ├── *.tsv
│     │     ├── *.meta
│     │     └── ...
│     └── [unique id]
├── requirements.txt  (requirements)
├── readme.md         (explaination)
├── losses.py         (useful losses)
├── metrics.py        (useful metrics)
├── model.py          (ESRGAN model)
├── main.py           (trainer / inferener)
├── config.py         (global configurations)
├── tfutils.py        (useful TF utils)
├── utils.py          (image processing utils)
└── dataloader.py     (DataSet loader)

Usage

  1. Clone this github repo.
git clone https://github.com/kozistr/ESRGAN-tensorflow
cd ESRGAN-tensorflow
  1. install required packages (if needed)
# with pip
python -m pip install -r requirements.txt

# with conda
conda install --yes --file requirements.txt
  1. run scripts!

For training,

python3 train.py

For evaluation,

python3 evaluate.py

For inference,

python3 inference.py --src test-lr.png --dst test-hr.png

Results

Citation

@InProceedings{wang2018esrgan,
    author = {Wang, Xintao and Yu, Ke and Wu, Shixiang and Gu, Jinjin and Liu, Yihao and Dong, Chao and Qiao, Yu and Loy, Chen Change},
    title = {ESRGAN: Enhanced super-resolution generative adversarial networks},
    booktitle = {The European Conference on Computer Vision Workshops (ECCVW)},
    month = {September},
    year = {2018}
}

Author

HyeongChan kim / kozistr

You can’t perform that action at this time.