DickJC123 commented on issue #14998: Revert the change on broadcast_to param 
shape
URL: https://github.com/apache/incubator-mxnet/pull/14998#issuecomment-495447354
 
 
   Thank you for your patience as I dealt with other MXNet fires.  I still feel 
I want to think about the Shape serialization to files.  But in the meantime, 
let me ask about the issue that started it all, and that you're addressing with 
this PR:
   
   Isn't the real problem that the Python output shape parameter of 
broadcast_to() isn't being converted correctly by the parameter parser?  In the 
default mode (so not numpy compatibility), a shape dimension size of 0 should 
be converted to -1 in the backend, which is what the broadcast_to() infer shape 
routine is already (i.e. correctly) looking for?  In numpy compatibility mode, 
the user should use -1 for the output shape in Python to denote unknown.

----------------------------------------------------------------
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