PyTorch: access weights of a specific module in nn.Sequential()

Question

When I use a pre-defined module in PyTorch, I can typically access its weights fairly easily. However, how do I access them if I wrapped the module in nn.Sequential() first? r.g:

class My_Model_1(nn.Module):
    def __init__(self,D_in,D_out):
        super(My_Model_1, self).__init__()
        self.layer = nn.Linear(D_in,D_out)
    def forward(self,x):
        out = self.layer(x)
        return out

class My_Model_2(nn.Module):
    def __init__(self,D_in,D_out):
        super(My_Model_2, self).__init__()
        self.layer = nn.Sequential(nn.Linear(D_in,D_out))
    def forward(self,x):
        out = self.layer(x)
        return out

model_1 = My_Model_1(10,10)
print(model_1.layer.weight)
model_2 = My_Model_2(10,10)

How do I print the weights now? model_2.layer.0.weight doesn't work.

Answer 1

An easy way to access the weights is to use the state_dict() of your model.

This should work in your case:

for k, v in model_2.state_dict().iteritems():
    print("Layer {}".format(k))
    print(v)

Another option is to get the modules() iterator. If you know beforehand the type of your layers this should also work:

for layer in model_2.modules():
   if isinstance(layer, nn.Linear):
        print(layer.weight)