Transfer learning using github repo

1

When we use pre-trained models from trochvision, we are using
model = models.vgg16(pretrained = True).
I am using a github repo of a network, which is not included in the torchvision. How can I assign the model?
model = XXXXXX(pretrained = True).
what should be this XXXXXX?

Any help would be appreciated. Sorry to ask a simple question,but I have been struggling with this.
Thanks

2

Hi, you need to simply instantiate the model and and load the state dict. Something like

model = Model() # create your model
model.load_state_dict(torch.load('model.pth')) #assuming you have the model weights saved in model.pth
4

Thanks a lot. I appreciate your help