Closed babakx closed 7 years ago
Hi! Thank you for contributing!
I believe it will be better to separate pred_batch_size
from __init__
and use it as param in decision_function
.
Also, I suggest setting pred_batch_size=None
by default and using pred_bach_size=batch_size
in this case.
Can you please fix this?
That's a good idea! I moved the parameter to the decision_function
and make the default value as None
just the way you explained.
Great, thanks!
Hi geffy! I did a minor change to make it possible to do the prediction in a different batch size than the batch size that is used for training.
This is useful when you want to predict the scores for a given matrix (for example test set matrix) in one go.
Best, Babak.