Cartoon-StyleGAN
Cartoon-StyleGAN copied to clipboard
Code to inference new image
Thank you for sharing this good work. I'm testing new image on the Colab inference notebook you have shared but unfortunately the Example 3 gives me bad results on all of images I uploaded. Don't know why although I run the exact same code on Colab notebook. Can you please take a look and share code for inference new images. Thanks in advance. I think it's because of encoder model. This is the image after run the projector.py line of code on Colab notebook.
Maybe that leads to the generated images being distorted too.
Hello,
You have to upload an image aligned with FFHQ method.
Re-arrange the image through the code of this repo (FFHQ-Alignment
) and encode the image.
This issue (#6) is example.
Thank you.