Skip to content

Latest commit

 

History

History

restoration

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
data
data
 
 
loss
loss
 
 
model
model
 
 
model_dhp
model_dhp
 
 
scripts/dhp
scripts/dhp
 
 
util
util
 
 
README.md
README.md
 
 
__init__.py
__init__.py
 
 
main.py
main.py
 
 
main_dhp.py
main_dhp.py
 
 

README.md

Dataset structure

├── super_resolution
│   ├── DIV2K
│   │   ├── DIV2K_train_HR
│   │   ├── DIV2K_train_LR_bicubic
│   │   │   ├── X2
│   │   │   ├── X3
│   │   │   ├── X4
│   │   ├── DIV2K_valid_HR
│   │   ├── DIV2K_valid_LR_bicubic
│   │   │   ├── X2
│   │   │   ├── X3
│   │   │   └── X4
│   │   └──
│   ├── benchmark
│   │   ├── Set5
│   │   ├── Set14
│   │   ├── B100
│   │   ├── Urban100
│   │   └──
│   └──
├── denoise
│   ├── DIV2KGRAY
│   │   ├── Train_HR
│   │   └──
│   ├── BenchmarkDenoise
│   │   ├── DenoiseSet68
│   │   │   ├── bin
│   │   │   └──
│   │   └──
│   └──
└──

Test

  1. Prepare image super-resolution and denoising dataset.

    i. Download DIV2K training and validation images in the NTIRE-2017 challenge.

    ii. Download super-resolution (Set5, Set14, B100, and Urban100) benchmark and denoising benchmark (Set68) from GoogleDrive or Dropbox.

    iii. Prepare the folder structure like the one above.

  2. Download the compressed models form Google Drive or Dropbox.

  3. cd ./scripts/dhp/

  4. Test the compressed DnCNN, UNet, SRResNet, and EDSR models.

            bash test_dncnn.sh
        bash test_unet.sh
        bash test_srresnet.sh
        bash test_edsr.sh

  5. Be sure to change the --pretrain, --dir_data, --dir_save directories.

Train

  1. cd ./scripts/dhp/

  2. Run the code to compress DnCNN, UNet, SRResNet, and EDSR networks.

    	bash dhp_dncnn.sh
	bash dhp_unet.sh
	bash dhp_srresnet.sh
	bash dhp_edsr.sh

  3. Be sure to change the --pretrain, --dir_data, --dir_save directories.