When I was playing around with slideflow on feature extraction, this line of code would always raise errors like ValueError: operands could not be broadcast together with shapes (4641,2048) (0,). This is because the left side could be either numpy array or tensor or just list (I guess). It would be good to check it in a more robust way. But I'm not sure if the modification in the PR is the best way to do it.
When I was playing around with slideflow on feature extraction, this line of code would always raise errors like
ValueError: operands could not be broadcast together with shapes (4641,2048) (0,)
. This is because the left side could be either numpy array or tensor or just list (I guess). It would be good to check it in a more robust way. But I'm not sure if the modification in the PR is the best way to do it.