I am currently trying to do binary classification, so I’ve set up a simple one input to 2 output neural network:
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.fc1 = nn.Linear(1,2)
def forward(self, x):
x = self.fc1(x)
return x
Using cross entropy loss, I attempt to classify inputs to a 0 or 1 labelling of data.
However, when attempting to train this model, there seems to be an issue when calculating a prediction:
for epoch in range(n_epochs):
for x_batch, y_batch in train_loader:
x_batch = x_batch.to('cpu')
print(x_batch.type())
y_batch = y_batch.to('cpu')
print(y_batch.type())
net.train()
yhat = net(x_batch.long()).unsqueeze(dim=0)
loss = loss_fn(yhat, y_batch)
loss.backward()
optimizer.step()
optimizer.zero_grad()
x_batch, and y_batch are both long, but when calculating x = self.fc1(x) in forward, this error comes up:
Traceback (most recent call last):
File "learned_bloom_filter.py", line 65, in <module>
yhat = net(x_batch).unsqueeze(dim=0)
File "/home/george/.local/lib/python3.6/site-packages/torch/nn/modules/module.py", line 541, in __call__
result = self.forward(*input, **kwargs)
File "learned_bloom_filter.py", line 16, in forward
x = self.fc1(x)
File "/home/george/.local/lib/python3.6/site-packages/torch/nn/modules/module.py", line 541, in __call__
result = self.forward(*input, **kwargs)
File "/home/george/.local/lib/python3.6/site-packages/torch/nn/modules/linear.py", line 87, in forward
return F.linear(input, self.weight, self.bias)
File "/home/george/.local/lib/python3.6/site-packages/torch/nn/functional.py", line 1372, in linear
output = input.matmul(weight.t())
RuntimeError: Expected object of scalar type Long but got scalar type Float for argument #2 'mat2' in call to _th_mm
I’ve searched around and found that I have to turn the weight of a Linear to scalar, but I’m not sure if this is the correct solution, nor do I know how to do it. Can someone advise?