You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Very clear and comprehensive work of ModernNCA! Respect to the team.
However, I have some questions about learning objective.
IIUC, during training, all of samples in candidate_x are necessary to be part of backward propagation. Thus, no matter what batch_size is set, the actual "batch_size" is the number of samples in candidate_x.
Is that right?
The text was updated successfully, but these errors were encountered:
Very clear and comprehensive work of ModernNCA! Respect to the team.
However, I have some questions about learning objective.
IIUC, during training, all of samples in candidate_x are necessary to be part of backward propagation. Thus, no matter what batch_size is set, the actual "batch_size" is the number of samples in candidate_x.
Is that right?
The text was updated successfully, but these errors were encountered: