Francesco comments

Results 87 comments of


                                            Francesco

documentation not working

Thanks for pointing it put. Yes, it is massively outdated, please refer to the readme for now. If then you still have any specific issues, let me know

Upscaling transparent PNGs

Hi, there is no way around flattening the image if you want to use the pretrained models. Look at this issue for alternative options: https://github.com/idealo/image-super-resolution/issues/158

Issue replicating the x4 RRDN weights from the sample weights

Hi, I recommend you first train using only MSE and then add the VGG loss and GANs. In my experiments I trained for about 60 epochs with MSE (PSNR) and...

Issue replicating the x4 RRDN weights from the sample weights

Yes, this is sth I have to fix. The issue is with monitored metrics: the name changes when no discriminator is used, so you have to change it accordingly, I...

Issue replicating the x4 RRDN weights from the sample weights

@Lorenzokrd does this happen after the initial MSE training phase too? How long did you train with MSE only? edit: I just recognized the 165 initial epochs of MSE. Still,...

Issue replicating the x4 RRDN weights from the sample weights

I cannot check the code right now, but I think you've been using the features from the VGG network too, not MSE only. If that's the case, this is kind...

InaccessibleTensorError: The tensor 'Tensor("Tile:0", shape=(None, 3), dtype=float32)' cannot be accessed here: it is defined in another function or code block.

Hi, thanks, that works. Quite clunky indeed. I did not realize that one of the issues was with the variable that defines the (outer) loop. Here are some examples of...