zixuanweeei opened a new pull request #17300: Align repr() function for rnn 
cell and rnn layer of Gluon
URL: https://github.com/apache/incubator-mxnet/pull/17300
 
 
   ## Description ##
   Align the repr() function of RNN cell with RNN layer in Gluon RNN API.
   
   
https://github.com/apache/incubator-mxnet/blob/58cbd65cfe98e50da36303c5f2adc65358d3a0b8/python/mxnet/gluon/rnn/rnn_layer.py#L121
   
   In RNN API, we often initialize the i2h weights with shape (`num_gates * 
hidden_dim`, `input_dim`). So `shape[0]` will get the size as hidden dimension 
times the number of gates. The repr() function should intend to print the 
information such as mapping the input dimension to the hidden/output dimension 
in the line above. In the RNN layers' API, it gets the right dimension, while 
it is not true in the cells' API.
   
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [x] Changes are complete (i.e. I finished coding on this PR)
   - [x] To the best of my knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   
   ### Changes ###
   - [x] Mapping input dim to hidden dim, instead of hidden dim times number of 
gates, for rnn cells.
   
   ## Comments ##
   + Is this issue critical for v1.6.x?
   
   @ciyongch @pengzhao-intel @TaoLv 
   
   

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


With regards,
Apache Git Services

Reply via email to