How can I change the backbone of the Deep Extreme Cut model from Resnet to say AlexNet or VGG19? The train_pascal file which is present in the PyTorch implementation of DEXTR seems to be calling resnet101.
How can I change the backbone of the Deep Extreme Cut model from Resnet to say AlexNet or VGG19? The train_pascal file which is present in the PyTorch implementation of DEXTR seems to be calling resnet101.