I’m currently having a problem I have trouble to solve.
I’m developing a CNN for text classification using pytorch but encounter this problem as I want to train my model :
loss, optimizer = createLossAndOptimizer(net, lr)
for epoch in range(num_epoch):
running_loss = 0.0 for i, data in enumerate(trainloader, 0): inputs, labels = data optimizer.zero_grad() outputs = net(inputs) loss = loss(outputs, labels) loss.backward() optimizer.step()
RuntimeError Traceback (most recent call last)
—> 16 loss = loss(outputs, labels)
/opt/conda/lib/python3.5/site-packages/torch/nn/functional.py in binary_cross_entropy(input, target, weight, size_average)
469 for each minibatch.
–> 471 return _functions.thnn.BCELoss(size_average, weight=weight)(input, target)
RuntimeError: expected a Variable argument, but got FloatTensor
both my outputs and my labels are a [torch.FloatTensor of size 64x1] type and dimension, which doesn’t seem to be working. I tried labels = Variable(labels) but it doesn’t work.
Can someone help me ?
Thank you very much.