study-faraphel/L1-SR-GUI
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
L1-SR-GUI/SRGAN-PyTorch-master
History
ℝλℙℍλΣʟ 50e4c082d8
Initial commit
2020-11-12 23:30:33 +01:00
..
__pycache__ Initial commit 2020-11-12 23:30:33 +01:00
dataset.py Initial commit 2020-11-12 23:30:33 +01:00
information.txt Initial commit 2020-11-12 23:30:33 +01:00
losses.py Initial commit 2020-11-12 23:30:33 +01:00
main.py Initial commit 2020-11-12 23:30:33 +01:00
mode.py Initial commit 2020-11-12 23:30:33 +01:00
ops.py Initial commit 2020-11-12 23:30:33 +01:00
readme.md Initial commit 2020-11-12 23:30:33 +01:00
srgan_model.py Initial commit 2020-11-12 23:30:33 +01:00
vgg19.py Initial commit 2020-11-12 23:30:33 +01:00

readme.md

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network

Overview

An unofficial implementation of SRGAN described in the paper using PyTorch.

Published in CVPR 2017

Requirement

  • Python 3.6.5
  • PyTorch 1.1.0
  • Pillow 5.1.0
  • numpy 1.14.5
  • scikit-image 0.15.0

Datasets

Pre-trained model

Train & Test

Train

python main.py --LR_path ./LR_imgs_dir --GT_path ./GT_imgs_dir

Test

python main.py --mode test --LR_path ./LR_imgs_dir --GT_path ./GT_imgs_dir --generator_path ./model/SRGAN.pt

Inference your own images

python main.py --mode test_only --LR_path ./LR_imgs_dir --generator_path ./model/SRGAN.pt

Experimental Results

Experimental results on benchmarks.

Quantitative Results

Method Set5 Set14 B100
Bicubic 28.43 25.99 25.94
SRResNet(paper) 32.05 28.49 27.58
SRResNet(my model) 31.96 28.48 27.49
SRGAN(paper) 29.40 26.02 25.16
SRGAN(my model) 29.93 26.95 26.10

Qualitative Results

Bicubic SRResNet SRGAN

Comments

If you have any questions or comments on my codes, please email to me. son1113@snu.ac.kr