oft icon indicating copy to clipboard operation
oft copied to clipboard

Published 20 hours ago • verified publisher icon tom-roddick

learning rate

Open jiayily opened this issue 3 years ago • 0 comments

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?

jiayily avatar Nov 02 '21 05:11 jiayily