/MDBN

Primary LanguagePythonMIT LicenseMIT

MDBN

Architecture

Multi-Depth Branches Network for Efficient Image Super-Resolution

Huiyuan Tian, Li Zhang, Shijian Li, Min Yao and Gang Pan

Run

Environments

  1. Install python3.10
  2. Install PyTorch (tested on Release 1.12)
  3. BasicSR 1.4.2

Installation

# Clone the repo
git clone https://github.com/thy960112/MDBN.git
# Install dependent packages
cd MDBN
pip install -r requirements.txt
# Install BasicSR
python setup.py develop

Training

For different scales, the following commands can be used for training respectively:

# train MDBN for x2 SR
python basicsr/train.py -opt options/train/MDBN/train_MDBN_x2.yml
# train MDBN for x3 SR
python basicsr/train.py -opt options/train/MDBN/train_MDBN_x3.yml
# train MDBN for x4 SR
python basicsr/train.py -opt options/train/MDBN/train_MDBN_x4.yml

Testing

  1. Download the pre-trained models.
  2. Download the testing dataset.
  3. For different scales, the following commands can be used for testing respectively:
# test MDBN for x2 SR
python basicsr/test.py -opt options/test/MDBN/test_MDBN_x2.yml
# test MDBN for x3 SR
python basicsr/test.py -opt options/test/MDBN/test_MDBN_x3.yml
# test MDBN for x4 SR
python basicsr/test.py -opt options/test/MDBN/test_MDBN_x4.yml

Acknowledgement

This code is based on BasicSR toolbox. Thanks for the awesome work.