lots of errors

[ngmisl]

❯ python main.py --im_path1 90.png --im_path2 15.png --im_path3 117.png --sign realistic --smooth 5
Loading StyleGAN2 from checkpoint: pretrained_models/ffhq.pt
Setting up Perceptual loss...
Traceback (most recent call last):
  File "main.py", line 117, in <module>
    main(args)
  File "main.py", line 18, in main
    ii2s = Embedding(args)
  File "/home/fg/dev/python/Barbershop/models/Embedding.py", line 25, in __init__
    self.setup_embedding_loss_builder()
  File "/home/fg/dev/python/Barbershop/models/Embedding.py", line 94, in setup_embedding_loss_builder
    self.loss_builder = EmbeddingLossBuilder(self.opts)
  File "/home/fg/dev/python/Barbershop/losses/embedding_loss.py", line 18, in __init__
    self.percept = lpips.PerceptualLoss(model="net-lin", net="vgg", use_gpu=use_gpu)
  File "/home/fg/dev/python/Barbershop/losses/lpips/__init__.py", line 22, in __init__
    self.model.initialize(model=model, net=net, use_gpu=use_gpu, colorspace=colorspace, spatial=self.spatial, gpu_ids=gpu_ids)
  File "/home/fg/dev/python/Barbershop/losses/lpips/dist_model.py", line 63, in initialize
    use_dropout=True, spatial=spatial, version=version, lpips=True)
  File "/home/fg/dev/python/Barbershop/losses/lpips/networks_basic.py", line 50, in __init__
    self.net = net_type(pretrained=not self.pnet_rand, requires_grad=self.pnet_tune)
  File "/home/fg/dev/python/Barbershop/losses/lpips/pretrained_networks.py", line 100, in __init__
    vgg_pretrained_features = tv.vgg16(pretrained=pretrained).features
  File "/home/fg/anaconda3/envs/plswork/lib/python3.7/site-packages/torchvision/models/vgg.py", line 150, in vgg16
    return _vgg('vgg16', 'D', False, pretrained, progress, **kwargs)
  File "/home/fg/anaconda3/envs/plswork/lib/python3.7/site-packages/torchvision/models/vgg.py", line 93, in _vgg
    progress=progress)
  File "/home/fg/anaconda3/envs/plswork/lib/python3.7/site-packages/torch/hub.py", line 559, in load_state_dict_from_url
    return torch.load(cached_file, map_location=map_location)
  File "/home/fg/anaconda3/envs/plswork/lib/python3.7/site-packages/torch/serialization.py", line 595, in load
    return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
  File "/home/fg/anaconda3/envs/plswork/lib/python3.7/site-packages/torch/serialization.py", line 781, in _legacy_load
    deserialized_objects[key]._set_from_file(f, offset, f_should_read_directly)
RuntimeError: unexpected EOF, expected 477241 more bytes. The file might be corrupted.

[Styml]

seems like you didn't download model properly and it was corrupted. Maybe try again?

[ngmisl]

seems like you didn't download model properly and it was corrupted. Maybe try again?

tried already. cloned the repo and made a clean conda env, results are the same, unfortunately.

[philippjess]

The model download happens when you run main.py, Your answer doesnt get to that step, for me gdrive is blocking the download so i had to do it manually.

[ngmisl]

Manually downloaded it too, as far as I'm concerned my setup is complete. 🙁

[philippjess]

sha256: 48a146c502117a283437d1c14f755a9416eb9255694e0964403f95b14fc094e9 ffhq.pt

[philippjess]

philipp@mars:~/git/ai/Barbershop/losses/lpips/weights/v0.1> sha256sum vgg.pth
a78928a0af1e5f0fcb1f3b9e8f8c3a2a5a3de244d830ad5c1feddc79b8432868  vgg.pth

This also gets downloaded in main.py