You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2018/03/28 20:45:38 UTC

[GitHub] ArmageddonKnight commented on issue #10304: Dropout has no effect in 1-Layer FusedRNNCell

ArmageddonKnight commented on issue #10304: Dropout has no effect in 1-Layer FusedRNNCell
URL: https://github.com/apache/incubator-mxnet/issues/10304#issuecomment-377030255
 
 
   Just to clarify, what the benchmark does seems to be the following:
   
   - For `range(num_layers)`, do the followings:
     - Declare one layer of `FusedRNN`.
     - Apply dropout to the `FusedRNN` output.
   
   However, why not do the following:
   
   - Declare `num_layers` of `FusedRNN` with `input_dropout` to be `dropout`.
   - Apply dropout to the final output (i.e. output at the last layer) of `FusedRNN`.
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services