Code for the GSRNet
tensorflow 1.4.0, python3.4, cuda 8.0.44 cudnn 6.0
Other packages please run:
pip install -r requirements.txt
Refer tohttps://github.com/DrSleep/tensorflow-deeplab-lfov for more detail
Run train_default.sh
- python3 dry_run.py --model_weights='./snapshots/$FIXME' --dataset='$FIXME'
for single image, use --dataset='single_img'
-
save output image: python3 dry_run.py --model_weights='./snapshots/$FIXME' --dataset='$FIXME' --save-dir='./output/$FIXME/' --vis=True --F1=True
-
visualize generated images: python3 dry_run.py --model_weights='./snapshots/$FIXME' --dataset='$FIXME' --save-dir='./output/$FIXME/' --vis=True --vis_gan=True