error rate compute in demo.py

[ghost]

Hi, thanks for this efficient densenet code.

But I found a probable mistake in demo.py at line 112:
error = 1 - torch.eq(predictions_var, target_var).float().mean()
it might have to be corrected to:
error = 1 - torch.eq(predictions_var.view(-1), target_var).float().mean()

Because the size of predictions_var and target_var are (train_size, 1) and (train_size, ), torch.eq(...) will return a train_size * train_size matrix, and its entries are almost 0 (only 1 at diagonal). Then the error rate will not able to decrease.

[taineleau-zz]

Hi, are you using PyTorch 0.2?

[ghost]

My PyTorch version is 0.2.0+925208a, I installing it from source.

[mosttl]

I met the same problem with the latest PyTorch (v0.2.0) installed, and also figured out the same solution.

[taineleau-zz]

The code currently is based on PyTorch v0.1.12. We will release the code for the latest PyTorch version soon. Please kindly use v0.1.12 for now. Thanks!

merged to #11