tqchen opened a new issue #8601:
URL: https://github.com/apache/tvm/issues/8601


   Followup of https://github.com/apache/tvm/issues/8579
   
   Right now our numerical r numerical gradient checker trying to do numerical 
gradient for all directions which scales linearly to the number of entries. 
   
   We can instead sample a few random directions to check so that the time 
complexity won't grow linearly to the number of entries. see page 8 of this 
[slide](https://catalyst.cs.cmu.edu/15-884-mlsys-sp21/slides/2-MLSys-AutomaticDifferentiation.pdf)


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to