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
In the paper you said "The model is trained using SGD for 600 epochs with a batch size of 8, momentum of 0.9 and learning rate of 10−7". Why we need to set learning rate so small?
The text was updated successfully, but these errors were encountered:
In the paper you said "The model is trained using SGD for 600 epochs with a batch size of 8, momentum of 0.9 and learning rate of 10−7". Why we need to set learning rate so small?
The text was updated successfully, but these errors were encountered: