PyTorch implementation of paper "Learned Scale-arbitrary Image Downscaling for Non-learnable Upscaling"
Download HR images:
Run:
# Ours + Bicubic
python run.py -model SAID_Bicubic -in_dir IMAGES_PATH -out_dir RESULTS_PATH -gpu 0 -scale 2.0
# Ours + Lanczos
python run.py -model SAID_Lanczos -in_dir IMAGES_PATH -out_dir RESULTS_PATH -gpu 0 -scale 2.0If you have any questions, please contact crhuang@whu.edu.cn.
