How to change the program to a classification model ?

[HuskyXPZ]

I have a time series data and I want to change the program to a model for classification.

The following is the part I changed in the program

Before：

# Output module
output = self._linear(decoding)
output = torch.sigmoid(output)

After：

# Output module
output = self._linear(decoding) 
output_softmax = torch.softmax(output)

After the modification, the program has the following error:

ValueError: Expected target size (64, 3), got torch.Size([64])

By the way, my data size is (64,55,20), and my category labels have three.

[maxjcohen]

Hi, we're going to need more information in order to better understand what is happening. What line raises this error, and the full backtrace ? What are your labels shape ? Please see #18 for more information about applying the model to a classification task.