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
Hi,
Sorry for the late reply. I found it achieve a good training speed - accuracy in the initial hyper-parameter search. It should have a close performance with 8 stacks 1 module.
I wonder why you set nModules(residual modules at each hourglass) to 2 rather than 1? Is there any special effect?
The text was updated successfully, but these errors were encountered: