Pytorch model that was used for the paper sent to ICOST 2020.
The model must be loaded like this:
# device is 'cuda' or 'cpu'. classes_names is an array with the classes names, but one could just pass
# a number instead of using len (I trained the model with 5 classes). Also be careful to change path/to/
# to something appropiate.
model = densenet121(channels=1, num_classes=len(class_names), drop_rate=0.7).to(device)
model = torch.nn.DataParallel(model).to(device)
model.load_state_dict(torch.load(path/to/cuda-alzheimer-densenet121.pth))