Umm what am I looking at?!?!?

taki0112 / UGATIT

Official Tensorflow implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation (ICLR 2020)

MIT License

6.17k stars 1.05k forks source link

Umm what am I looking at?!?!? #111

Open Pleasehelo opened 3 years ago

Pleasehelo commented 3 years ago

So I wanted to see what my daughter looked like in anime but it honestly looked ugly. here is a photo of her: IMG_20210130_131948 Anime:

sxf97 commented 3 years ago

It's pretty good, but I don't think the training is very good

tlbjr commented 3 years ago

It seems really close to being excellent. I don't know a thing about image processing or training, but maybe adjust the color balance because the window light is dominating all the subtle face lighting and you probably want more of a portrait look.

guchengxi1994 commented 3 years ago

why my results are such a mess?? but the test images performed well 企业微信截图_16152570597616 企业微信截图_16152571267760

syz825211943 commented 2 years ago

@guchengxi1994 Have you found the solution? I met this problem too. The model cannot work well on my own dataset but can produce well results on its own dataset.

guchengxi1994 commented 2 years ago

@syz825211943 i was using ugatit pytorch version https://github.com/znxlwm/UGATIT-pytorch and i think that is better than this.
I used my own dataset and made a funny thing. https://github.com/guchengxi1994/Hand-painted-faces-to-real-faces