your notebook is fast and fun re: CLIP-guided art on faces. Do you have a suggestion on where to look for PKL files other than FFHQ? I've tried Justin Pickney's site, but when I try one of his pkl files with your code by creating a new network_pkl path:
with dnnlib.util.open_url(network_pkl) as f: G = legacy.load_network_pkl(f)['G_ema'].to(device) # type: ignore
I get a variety of errors depending on the pkl file (such as "unpickling" errors). I am guessing his pkl files are in a different format, although I thought they were all TF pickle files.
I also posted an issue on his repository, which is also in GitHub and I will make sure to cross-connect anything I learn in this process.
Thank for for making this notebook.
your notebook is fast and fun re: CLIP-guided art on faces. Do you have a suggestion on where to look for PKL files other than FFHQ? I've tried Justin Pickney's site, but when I try one of his pkl files with your code by creating a new network_pkl path:
with dnnlib.util.open_url(network_pkl) as f: G = legacy.load_network_pkl(f)['G_ema'].to(device) # type: ignoreI get a variety of errors depending on the pkl file (such as "unpickling" errors). I am guessing his pkl files are in a different format, although I thought they were all TF pickle files.
I also posted an issue on his repository, which is also in GitHub and I will make sure to cross-connect anything I learn in this process.
Thank for for making this notebook.