szha commented on issue #10453: Bug of CuDNN RNN with variable sequence length
   What I observed is that it doesn't fail consistently on certain specific 
batch. Another team observed the same issue before, and it is likely caused by 
our backend memory pool holding too much memory, in which case the curand 
doesn't have enough memory to keep the random number generator states for each 
stream multiprocessor.

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:

With regards,
Apache Git Services

Reply via email to