yunzhe-tao opened a new issue #11253: CUDNN implementation of LS
URL: https://github.com/apache/incubator-mxnet/issues/11253
 
 
   Hi, I want to implement the LSTM projection layer on MXNet Gluon. I was 
trying to implement the LSTMP layer similar to gluon.rnn.lstm based on 
gluon.contrib.rnn.LSTMPCell but I found that the ndarray.rnn function in 
LSTM._forward_kernel is implemented based on C++ and the codes are not 
available. Moreover, the function only works for RNN, LSTM and GRU. Without 
that function, I found LSTMP layer runs really slow. So I am wondering if 
someone can implement _forward_kernel for more general recurrent layers. Thanks.

----------------------------------------------------------------
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:
us...@infra.apache.org


With regards,
Apache Git Services

Reply via email to