Skip to content

Latest commit

 

History

History
66 lines (53 loc) · 1.64 KB

README_zh_cn.md

File metadata and controls

66 lines (53 loc) · 1.64 KB

代码运行文档

复现代码

本代码根据论文《Deep Photo Enhancer: Unpaired Learning for Image Enhancement from Photographs with GANs》。

原作者代码库：nothinglo/Deep-Photo-Enhancer

所需环境

  • Python 3.6
  • CUDA 10.0
  • 具体见requirements.txt，使用如下命令安装依赖： pip install -r requirements.txt

所需数据集

Expert-C on MIT-Adobe FiveK dataset

部分操作

  1. PreTrain和Train中的参数可以修改
  2. 在SourceCode下要新建多个文件夹，分为：
    1. images_LR：用来存放数据集
      1. Expert-C
      2. input
      3. 上述两文件夹中都需要新建下面三个文件夹
        1. Testing
        2. Training1
        3. Training2
    2. model：用来存放所有训练产生的文件，其下还需新建：
      1. gt_images
        1. 1Way
        2. 2Way
      2. input_images
        1. 1Way
        2. 2Way
      3. pretrain_checkpoint
        1. 1Way
        2. 2Way
      4. pretrain_images
        1. 1Way
        2. 2Way
      5. test_images
        1. 1Way
        2. 2Way
      6. train_checkpoint
        1. 1Way
        2. 2Way
      7. train_images
        1. 1Way
        2. 2Way
      8. train_test_images
        1. 1Way
        2. 2Way
    3. models:用来存放log_PreTraining.txt
    4. 训练后的gan1_pretrain_XXx_xxx.pth需要放在根目录下

Cost Time

  • Pretrain: 3H55M 8H45M 9H25M
  • Train: 2H45M 2H49M 3H03M 5H38M(2Way) 4H45M(2Way)

Problem

  1. There may be a problem in computing the value of PSNR or not. It needs to be proved.