Shape mismatch between task output and task _pred output

[atapley]

After running train.py and retrain.py, I run test.py and am met with some errors in a shape mismatch.

I am trying to get the precision_score using the sklearn.metrics package, but the task output from the GT is of shape (,) while the task _pred output is of shape (1). To fix this issue, I have to make a change in the task's error method in base_env.py to use torch.unsqueeze.

Any idea why there would be a shape mismatch all of a sudden after having worked for both train.py and retrain.py?