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
I am using the SVD++ algorithm. I found that the process of training is hardly coded and the weight will be re-initialized before training once we call the fit for the second time. So I cannot come up with any ideas to do early stopping with official surprise. Could we add support for early stopping, or is there some reason for not adding it ?
The text was updated successfully, but these errors were encountered:
Hi @yslking , sorry for the late reply. I don't have as much time as I used to, so I'm mostly just maintaining the project now; I'm unable to commit to new features, sorry
I am using the SVD++ algorithm. I found that the process of training is hardly coded and the weight will be re-initialized before training once we call the
fit
for the second time. So I cannot come up with any ideas to do early stopping with official surprise. Could we add support for early stopping, or is there some reason for not adding it ?The text was updated successfully, but these errors were encountered: