CUDA_ERRIR_OUT_OF_MEMORY about stylegan2 HOT 4 OPEN

11GB is probably too small depending on what you’re training, especially is you’re running additional processes on it.

from stylegan2.

here is GPU information of my second computer, I had no luck running your repository on both of them, you replied to my youtube comment by stating that I should have GPU=16GM!, is the below specifications not enough

from stylegan2.

I see, thank you very much @dvschultz I have another question

how can I create stylegan model with (1, 18, 512)

my stylegan model is creating shape (1, 12, 512) and I cannot find the latent space developed by Puzer because of shape difference

in more details:
my model produce shape with (1, 12, 512) using (https://github.com/NVlabs/stylegan)
but when I use stylegan encoder (https://github.com/Puzer/stylegan-encoder) to find latent space it requires (1, 18, 512), do you have any idea how can I produce (1, 18, 512) model shapes instead of (1, 12, 512)?

from stylegan2.

what size output is your model? As I recall only 1024 does (1,18,512). Smaller resolutions will generate smaller shapes. Many of the encoders are only set up to with with FFHQ and its 1024^2 resolutions.

from stylegan2.