Awesome

Lightweight multi-scale distillation attention network for image super-resolution

[python 3.8]

[Ubuntu 20.04]

git clone https://github.com/Supereeeee/MSDAN.git
pip install -r requirements.txt
python setup.py develop

· Refer to ./options/test for the configuration file of the model to be tested and prepare the testing data.

· The pre-trained models have been palced in ./experiments/pretrained_models/

· Then run the follwing codes (taking MSDAN_x4.pth as an example):

python basicsr/test.py -opt options/test/test_MSDAN_x4.yml

The testing results will be saved in the ./results folder.

· Refer to ./options/train for the configuration file of the model to train.

· Preparation of training data can refer to this page. All datasets can be downloaded at the official website.

· Note that the default training dataset is based on lmdb, refer to docs in BasicSR to learn how to generate the training datasets.

· The training command is like

python basicsr/train.py -opt options/train/train_MSDAN_x4.yml

For more training commands and details, please check the docs in BasicSR

· The network structure of MSDAN is palced at ./basicsr/archs/MSDAN_arch.py

· We adopt thop tool to calculate model complexity, see ./basicsr/archs/model_complexity.py

· We test the inference time on multiple benchmark datasets on a 140W fully powered 3060 laptop.

· You can run ./inference/inference_MSDAN.py on your decive.

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

If you have any question, please email 1051823707@qq.com.