How to change image size?

[armeausstahl]

Hey @ZPdesu ,

I love your project and currently working on some test cases. I want to improve the speed of embeddings by using smaller pictures. Is there a easy to achieve this?

I used
python align_face.py -output_size 512
to generate the pictures with size 512

after running
python main.py --im_path1 90.png --im_path2 20.png --im_path3 20.png --sign realistic --smooth 5 --size 512

I got this error:

RuntimeError: Error(s) in loading state_dict for Generator:
        Unexpected key(s) in state_dict: "convs.14.conv.weight", "convs.14.conv.blur.kernel", "convs.14.conv.modulation.weight", "convs.14.conv.modulation.bias", "convs.14.noise.weight", "convs.14.activate.bias", "convs.15.conv.weight", "convs.15.conv.modulation.weight", "convs.15.conv.modulation.bias", "convs.15.noise.weight", "convs.15.activate.bias", "to_rgbs.7.bias", "to_rgbs.7.upsample.kernel", "to_rgbs.7.conv.weight", "to_rgbs.7.conv.modulation.weight", "to_rgbs.7.conv.modulation.bias", "noises.noise_15", "noises.noise_16".

[ZPdesu]

I did not do tests on 512 x 512 images. Maybe you should download the weights of the StyleGAN generator trained on 512x512 FFHQ images.

[deng-wei]

I did not do tests on 512 x 512 images. Maybe you should download the weights of the StyleGAN generator trained on 512x512 FFHQ images.

I also encountered this problem. I want to use a smaller size of StyleGAN2 due to the limitation of the machine resources. But I don't know how to get the .pt file. Are you trained by yourself or converted from the official .pkl weights?

[sunzongzheng]

same issue here