[GitHub] reminisce opened a new pull request #7244: Operator add_n for row sparse ndarrays

[gi...@git.apache.org]

reminisce opened a new pull request #7244: Operator add_n for row sparse ndarrays
URL: https://github.com/apache/incubator-mxnet/pull/7244
 
 
   1. Added [add_n](http://mxnet.io/api/python/ndarray.html?highlight=add_n#mxnet.ndarray.add_n) operator for row sparse ndarrays.
   2. Added IdentityComputeEx for dense/sparse ndarrays as FComputeEx, which is required in the backward pass of add_n operator. This work may overlap with @cjolivier01 's, but I need this function to make add_n backward pass work for row sparse inputs.
   3. Fixed the bug of not filling zeros for the output of square_sum forward FComputeEx function when the input sparse ndarray is not storage_initialized in this PR, https://github.com/apache/incubator-mxnet/pull/7206
   4. Removed test_cast_storage_ex from test_operator_gpu.py since it has not been completely implemented.
   
   @piiswrong @eric-haibin-lin @madjam @cjolivier01 @anirudh2290 @stefanhenneking 
   
 
----------------------------------------------------------------
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