Francesco
Francesco
Are the images you're loading all RGB (3 channels)? I have not seen this before. And thanks for the suggestion, there should definitely be an issue template.
@MCOfficer you are correct, there should be a channel consistency check for the training set, at the moment the only checks are at prediction time. Thank you for you suggestion!
I'm not sure what happens there. The model is asking for a 4 dimensional input (a batch of images) but it's receiving an array. The code you posted seems correct....
> I am training it on CelebA Dataset. I cropped all the faces from the dataset to size 128x128 and the LR images are 32x32 upscaled to 128x128 using bicubic...
This question is a little too vague, please be more specific on what issues you're encountering.
so it seems you're using only the gan loss. I would recommend to first train with the psnr loss and see the results afterafter 30-40 epochs. If that is going...
Hi, generator is the loss weights for the plain network, having this to 1 and the others to 0 will only do pixelwise MAE with the original image. feature_extractor is...
Hi, no, only using mae you will get less defined pictures. You can try adding only the feature loss, if training with GANs is too difficult
Hi, a PR for a flask interface would certainly be most appreciated! I'm not sure what you are asking for here, no rescaling of the input picture or no you...
Hi, maybe then you should look more in the direction of image restoration. However, setting the scale parameter ('x') would allow you to use this repo to train a network...