Tamar Rott Shaham
Tamar Rott Shaham
yes, you can adjust max_size according to the desire output size.
yes, but you may encounter some computation limitation
you need to use `max_size` instead of `max-size`
All the SR evaluation were done using the code of the work "the perception-distortion tradeoff" by Blau et. al. Code is available at their project webpage: https://webee.technion.ac.il/people/tomermic/PerceptionDistortion/PD_tradeoff.htm
The reported numbers are average over the full BSD100 dataset (not per a single result). You can find all the results in the "Downloads" folder.
Thanks. We intentionally did it this way. This is because we refer the *patches* of the image as our data-set, and we want the discriminator to be 1-lipschitz per patch....
Hi, You can transfer it to be a loss over the whole image by flattening the data: ``` gradients = gradients[0].view(real_data.size(0), -1) ``` before calculating the penalty itself. The whole...
In all the scales we use a patch-discriminator which has a small receptive field of 11x11, therefore the discriminator never sees the full image, but only 11x11 patches from it,...
the function calculates the scale factor closest as possible to --scale_factor such that the image at the coarsest scale will be of size --min_size and at the finest scale of...
you can create the naively pasted version and the mask manually, using photoshop/paint.net or similar programs