is it possible to change it in config witout affecting the performance?
Hey, this is something you have to train into the model. But you can just oversample if you have a smaller corpus; that should work fine.
· Sign up or log in to comment
