Please create the full model (weights plus architecture) to torch.hub.load

[spinoza1791]

Thank you for your amazing work! If your team is able to upload the full model (weights plus architecture) to torch.hub.load then the model could be converted to a mobile version. Or do you already have the pretrained model weights plus architecture file? I can see that your .pth is only the OrderedStrict weights.

[ptran1203]

Unfortunately, only model weights are available at the moment. But you can easily derive it with some Python code (I did not check the syntax)

import torch
from modeling.anime_gan import Generator

weight = "..."
model = Generator()
checkpoint = torch.load(weight,  map_location='cpu')
model.load_state_dict(checkpoint['model_state_dict'], strict=True)
model.save(model, "/path/to/full_model.pth")

[spinoza1791]

Understood, but could you give an example of what to enter for these two suggested lines, as I cannot figure out what to enter for them? Thank you!

weight = "..."

model.load_state_dict(checkpoint['model_state_dict'], strict=True)

[ptran1203]

weight = "..." is path to model weight on your machine. something like this weight model.load_state_dict(checkpoint['model_state_dict'], strict=True) to get model state_dict, just keep as it is, no need to change

[spinoza1791]

Thank you, I am still getting errors however on this line: model.load_state_dict(checkpoint['model_state_dict'], strict=True)

Any ideas?

import torch from torch import Generator

weight = 'anime.pth' model = Generator() checkpoint = torch.load(weight, map_location=torch.device('cpu')) model.load_state_dict(checkpoint['model_state_dict'], strict=True) model.save(model, "anime_sd.pth")

----> 7 model.load_state_dict(checkpoint['model_state_dict'], strict=True) 8 model.save(model, "anime_sd.pth")

AttributeError: 'torch._C.Generator' object has no attribute 'load_state_dict'