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
{{ message }}
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
if we fix the parameters of an operator in the proceeding of training, (that means we do not learn the parameters), then its operator should be in an "is_train=false" mode. the "forward" should use the inference code in it.
The text was updated successfully, but these errors were encountered:
starimpact
changed the title
advise:set "is_train" of operator be "false" when its parameter gradient is null
advise:set "is_train" of operator be "false" when its gradient parameter is null
Nov 7, 2017
@starimpact Could you please add more information on why this is a problem and add small code to reproduce the problem? How did you freeze the parameter? Are you using the symbol or the imperative API?
if we fix the parameters of an operator in the proceeding of training, (that means we do not learn the parameters), then its operator should be in an "is_train=false" mode. the "forward" should use the inference code in it.
The text was updated successfully, but these errors were encountered: