Training with high resolution cartoon images

[menu23]

Thank you for your wonderful repository.
My question is, I want to use cartoon images that have very high resolution and detail (2K).
Is it a good idea to split/slice the images into smaller pieces (2x2 or 4x4) before training to get better results?
Any other suggestions are welcome.

[WynMew]

Thank you for your wonderful repository.
My question is, I want to use cartoon images that have very high resolution and detail (2K).
Is it a good idea to split/slice the images into smaller pieces (2x2 or 4x4) before training to get better results?
Any other suggestions are welcome.

I've tried that. The key here is not the cartoon images resolution, but the correspondence between cartoon image 'content' and the real image 'content'. i.e. if you slice cartoon images into small pieces, you get local content(parts of trees, buildings and bodies) instead of 'global content'(the full story). Then you lost the correspondence of 'content' between your cartoon training images and real images, and you lost the style correspondence and get always big 'style loss'.