Simon Karlsson

Hi, The default initialization of the weights is different for different layers. Read about each layer in the keras documentation. In section 'Architecture functions' from row 247 the weights are...

Hi @MrtnStnwk, Thank you for your find, I will update the model right away.

Hi @MrtnStnwk, Seemed to have missed your last post. Great that you are commenting the errors, thank you. I remember working on the error with number of samples needing to...

Thank you for the assistance @kpagels! Are there any translated synthetic images in any of your new folders `generated_images/synthetic_images/A` or `generated_images/synthetic_images/B` @HeoJinLareine?

Hi @danielJang93, Try using the option of loading the images during training. Enable this on row 75 in the model by `self.use_data_generator = True`

Hi @adilincepto, Without the adversarial loss the model would perform very badly. You should read up a bit more to see all parts included. First note that the variable 'compile_losses'...

Hi @HeoJinLareine, I have no experience in this unfortunately. I hope you can figure it out even though I'm not much help! Regards, Simon

Hi @YINDAIYING, The research project was done in collaboration with a private company that provided the dataset and which has the right to all used training and test images. Thereby...

No worries, hope you find what you are looking for :) A recommendation is that you take a look at the original CycleGAN repository where you can find and download...

Hi @SZUHvern, After some googling I see that Keras updates have made it easier to use multiple GPUs. I suggest you look into if you need to modify anything in...