Hi, I was working with CIFAR-10 dataset and Resnet18 model. I am using inbuilt model architecture and calling it as
net = torchvision.models.resnet18(pretrained=True)
I am using ImageNet weights and training the entire model. I wanted to add dropout to the network after various layers. I don’t know how to add dropout to the inbuilt network called. And the network summary shows there’s only one FC(or linear) layer in the network in the end. So where can we add dropout in the network? As per my knowledge, isn’t it supposed to be added to the FC layers in the network?
Any suggestions around that?
Thanks!