I’m trying to penalize the norm of activation gradients:
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(3, 32, kernel_size=5)
self.conv2 = nn.Conv2d(32, 64, kernel_size=5)
self.pool = nn.MaxPool2d(2, 2)
self.relu = nn.ReLU()
self.linear = nn.Linear(64 * 5 * 5, 10)
def forward(self, input):
conv1 = self.conv1(input)
pool1 = self.pool(conv1)
self.relu1 = self.relu(pool1)
self.relu1.retain_grad()
conv2 = self.conv2(relu1)
pool2 = self.pool(conv2)
relu2 = self.relu(pool2)
self.relu2 = relu2.view(relu2.size(0), -1)
self.relu2.retain_grad()
return self.linear(relu2)
model = Net()
optimizer = torch.optim.SGD(model.parameters(), lr=0.001)
for i in range(1000):
output = model(input)
loss = nn.CrossEntropyLoss()(output, label)
optimizer.zero_grad()
loss.backward()
grads = torch.autograd.grad(loss, [model.relu1, model.relu2], create_graph=True)
grad_norm = 0
for grad in grads:
grad_norm += grad.pow(2).sum()
grad_norm.backward(retain_graph=True)
optimizer.step()
However, it does not produce the desired regularization effect. If I do the same thing for weight gradients, it works well. Am I doing this right? Specifically, what happens in grad_norm.backward()
step? I just want to make sure the weight gradients are updated, and not activation gradients. Currently, when I print out gradients for weights and activations immediately before and after that line, both change - so I’m not sure what’s going on.