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
Currently update_batch_norm just runs the librispeech workloads in train mode, which also runs dropout in train mode. The purpose of having separate mode and update_batch_norm kwargs to model_fn() was so that submitters could separate which they want to update, if desired. We can update Conformer.__call__ and Deepspeech.__call__ to take both train and update_batch_norm and pass them to dropout/BN respectively.
The text was updated successfully, but these errors were encountered:
Currently
update_batch_norm
just runs the librispeech workloads in train mode, which also runs dropout in train mode. The purpose of having separatemode
andupdate_batch_norm
kwargs tomodel_fn()
was so that submitters could separate which they want to update, if desired. We can updateConformer.__call__
andDeepspeech.__call__
to take bothtrain
andupdate_batch_norm
and pass them to dropout/BN respectively.The text was updated successfully, but these errors were encountered: