search

MIVRC / MSRN-PyTorch

This repository is a PyTorch version of the paper "Multi-scale Residual Network for Image Super-Resolution" (ECCV 2018).
MIT License
292 stars 56 forks source link

Testing on patches #26

Closed harish-vnkt closed 4 years ago

harish-vnkt commented 4 years ago

Hi, during testing on the benchmark datasets, do you use the low resolution patches or do you use the whole image?

CV-JunchengLi commented 4 years ago

We use the whole image during testing.

In order to reduce memory and training time, the patch strategy is used during training, but not used during testing.

harish-vnkt commented 4 years ago

Thank you! I have another question. Is it possible that some of the pixels of an image might not get selected in a patch during the entire training phase?

CV-JunchengLi commented 4 years ago

It's possible. The strategy of randomly selecting patches is used during training, so there is such a possibility.