Hi I am pretty new to PyTorch and neural networks in general and I am wondering about the weight vector estimate of linear networks. Suppose I have built a very simple model with 3 input nodes and 2 output nodes (no hidden layers) to use on a 2D dateset I generated (I know what the weight vector should look like). I only have two different classes in this set and I would like to extract the weight vector (linear decision boundary) from my network once training is done. I was wondering what is the best way to do that.
So far I have tried something like this :
weights_out = list(model.fc1.weight.data)
which gives me a pair of 2D vectors, one for each output node I presume.
Thanks in advance for any help.
My network actually looks like this :
class Net(nn.Module):
def __init__(self,dimensions,model_type):
super(Net,self).__init__()
self.fc1 = nn.Linear(3,2)
def forward(self, x):
x = self.fc1(x)
return x