[GitHub] [incubator-mxnet] eric-haibin-lin commented on issue #15215: MXNet Row Sparse SGD Update fails when Row Sparse weights are sparsely populated

[GitBox <gi...@apache.org>]

eric-haibin-lin commented on issue #15215: MXNet Row Sparse SGD Update fails when Row Sparse weights are sparsely populated
URL: https://github.com/apache/incubator-mxnet/issues/15215#issuecomment-503422954
 
 
   For this particular operator, I think it's possible just to enhance the operator with a kernel that handles true sparse weight and sparse gradient. 
   
   Can we assume the non-zero indices appeared in the grad can always be found in the indices of the weight? If so that makes it easier to implement. There's a guide for data structure related to sparse ndarray in https://cwiki.apache.org/confluence/display/MXNET/A+Guide+to+Implementing+Sparse+Operators+in+MXNet+Backend and I'm happy to give more pointers if needed. 

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services