CoarseNet

CoarseNet copied to clipboard

VGG part of loss function cannot be decreased at all.

1

The generated output of the model regarding only this loss, is a full noise even after unnormalization. - [ ] Write a function to create a feature map of contributed...

Nikronic

Use this approach: https://discuss.pytorch.org/t/visualize-feature-map/29597/2

Nikronic

Mean and STD values regarding our dataset is required

3

We need to calculate `mean` and `std` of the whole dataset lazily (online) BEFORE applying `halftone`. Actually, The output is something similar to ground truth and by the way we...

Nikronic

CoarseNet is not working as a low-pass filter - Halftone patterns are visible

1

Training loss cannot be declined from 30 regarding default weights.

Nikronic