Closed Sandeep9343 closed 1 year ago
Hello,
I'm encountering a couple of issues while using this code for training on a custom dataset for watermark removal:
After 200 epochs, the 'D_fake' value reaches zero, which leads to the generation of a blank image in the evaluation folder.
When using the 'gen.pth.tar' model for prediction, it's producing a blank image.
Could you please provide some assistance with these problems?
This code is written by looking on the YouTube video
i took the code from github still getting nan values while training
Hello,
I'm encountering a couple of issues while using this code for training on a custom dataset for watermark removal:
After 200 epochs, the 'D_fake' value reaches zero, which leads to the generation of a blank image in the evaluation folder.
When using the 'gen.pth.tar' model for prediction, it's producing a blank image.
Could you please provide some assistance with these problems?