Weird results

[ofirkris]

Thanks for sharing this! I've tested with the pre-trained models, inserted them to the "checkpoint" folder Ran this command:

python main.py --dataset FFHQ --img_size 1024 --progressive True --phase draw --draw uncurated --checkpoint_dir checkpoint

And got the following output + the attached results:

" Total size of variables: 26219627 Total bytes of variables: 104878508 [] Reading checkpoints... [] Failed to find a checkpoint [!] Load failed... [*] Un-curated finished! "

What am I doing wrong? Also, how can I change the input image?

[enric1994]

I got the same output. Try running the same with --progressive False

[ofirkris]

Hi @enric1994 thanks, I've tried --progressive False parameter, it still doesn't work. I think it does not work because the loading of the checkpoint failed but I don't understand why..

[taki0112]

@ofirkris @enric1994

load failed means that the checkpoint could not be called. That is, maybe you didn't put the checkpoint file I provided into the proper path.

1. Download the checkpoint
2. unzip
3. put the checkpoint like this : ./checkpoint/StyleGAN_FFHQ_8to1024_progressive/checkpoint ./checkpoint/StyleGAN_FFHQ_8to1024_progressive/StyleGAN.model-224999.data-00000-of-00001 ./checkpoint/StyleGAN_FFHQ_8to1024_progressive/StyleGAN.model-224999.index ./checkpoint/StyleGAN_FFHQ_8to1024_progressive/StyleGAN.model-224999.meta

[ofirkris]

Thanks @taki0112 ! finally - How can I change the input image?

[taki0112]

@ofirkris I don't understand what you mean. What do you mean by changing the input image?

[ofirkris]

From what I understand the input images you're using are (StyleGAN.py line 642): src_seeds = [604, 8440, 7613, 6978, 3004] dst_seeds = [1336, 6968, 607, 728, 7036, 9010]

What if I want to test on my own image?

[taki0112]

If you want to your image, you need to obtain a latent vector for the image. Please refer to this