• Python 3.8
• PyTorch 1.8.0
• NVIDIA GPU + CUDA

# Clone the github repo and go to the default directory 'codebase_SR'.
git clone https://github.com/yulunzhang/codebase_SR
cd codebase_SR
conda create -n sr_pytorch python=3.8
conda activate sr_pytorch
pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio==0.8.0 -f https://download.pytorch.org/whl/torch_stable.html
pip install -r requirements.txt
python setup.py develop

After training starts or ends, view the validation (or loss) curve through Tensorboard.

Here, we take the example of training a model on a server and viewing Tensorboard locally.

All Tensorboard-log files are in tb_logger.

Run the following scripts in the server.

ssh -p 22 -L 16006:127.0.0.1:6006 username@remote_server_ip
tensorboard --logdir=tb_logger/xxx --port=6006

Then open the web page http:https://localhost:6006/ on your local computer.

1. Datasets
2. Models
3. Training
4. Testing

Used training and testing sets can be downloaded as follows:

Download training and testing datasets and put them into the corresponding folders of datasets/. See datasets for the detail of the directory structure.

The performance is reported on Urban100, and output size of FLOPs is 3×1280×720.

We also provide the testing results (log) and tb_logger in the folder.

• Download training (DF2K, already processed) and testing (Set5, Set14, BSD100, Urban100, Manga109, already processed) datasets, place them in datasets/.

• Run the following scripts. The training configuration is in options/train/.

  # FSRCNN, x2, input=48x48, 1 GPU
  python basicsr/train.py -opt options/Train/train_FSRCNN_x2.yml
  
  # FSRCNN, x3, input=48x48, 1 GPU
  python basicsr/train.py -opt options/Train/train_FSRCNN_x3.yml
  
  # FSRCNN, x4, input=48x48, 1 GPU
  python basicsr/train.py -opt options/Train/train_FSRCNN_x4.yml

• The training experiment is in experiments/.

• Run the following scripts. The testing configuration is in options/test/ (e.g., test_FSRCNN_x4.yml).

  # FSRCNN, x2
  python basicsr/test.py -opt options/Test/test_FSRCNN_x2.yml
  
  # FSRCNN, x3
  python basicsr/test.py -opt options/Test/test_FSRCNN_x3.yml
  
  # FSRCNN, x4
  python basicsr/test.py -opt options/Test/test_FSRCNN_x4.yml

• The output is in results/.

• Put your dataset (single LR images) in datasets/single. Some test images are in this folder.

• Run the following scripts. The testing configuration is in options/test/ (e.g., test_single_x4.yml).

  # Test on your dataset, x2
  python basicsr/test.py -opt options/Test/test_single_x2.yml
  
  # Test on your dataset, x3
  python basicsr/test.py -opt options/Test/test_single_x3.yml
  
  # Test on your dataset, x4
  python basicsr/test.py -opt options/Test/test_single_x4.yml

• The output is in results/.