Implementation of super-resolution using DRRN.
-Python 3.6
-PyTorch 1.5
- download the data
- the training data is generated with Matlab Bicubic Interpolation, please refer Code for Data Generation for creating training files.
-run the following command to train:
python train.py --dataset (the dataset you want to train) --outdirname (the output directory)
-run the following command to generate sr-pictures:
python sr.py --model model_7000.npz --image x.png, x is the number of the picture