search

junyanz / pytorch-CycleGAN-and-pix2pix

Image-to-Image Translation in PyTorch
Other
22.8k stars 6.29k forks source link

How to manipulate or create a new loss function ?? #1194

Open shivom9713 opened 3 years ago

shivom9713 commented 3 years ago

Dear sir , I am working on Infrared to visual image conversion using GANs , unfortunately cycle consistency loss and adversarial loss is not enough to make good conversions even after good amount of training on a sufficiently large dataset . Can you help me what loss functions should i use and how to make changes in loss function code ?? And also how can we do semi-supervised learning using cycle GANs Thanks

seawee1 commented 3 years ago

Hi, there are many follow-up works based on CycleGAN which might be better suited to tackle your problem, probably also the semi-supervised setting. I can very much recommend this repository to look up some alternative methods: Awesome Image Translation

shivom9713 commented 3 years ago

Thanks a lot sebastian !!

On Wed, Nov 25, 2020 at 7:06 PM Sebastian Rietsch notifications@github.com wrote:

Hi, there are many follow-up works based on CycleGAN which might be better suited to tackle your problem, probably also the semi-supervised setting. I can very much recommend this repository to look up some alternative methods: Awesome Image Translation https://github.com/weihaox/awesome-image-translation

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/1194#issuecomment-733706984, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANNEXP4VA64JDKTAT3S6SI3SRUBFFANCNFSM4T4JLDPA .