jimmychou0704 commented on issue #13748: custom loss cause "kernel died" in 
Jupyter notebook
URL: 
https://github.com/apache/incubator-mxnet/issues/13748#issuecomment-450909728
 
 
   @wkcn 
   Thanks, you are right. The problem is shapes of inputs and outputs of 
backward operation don't match. Once I changed my code to 
   `self.assign(in_grad[0], req[0], (out_data[0] - in_data[1].reshape((-1,1)) 
))`
   the issue is gone.
   I suppose the formula is correct, i followed the discussion in 
http://cs231n.github.io/neural-networks-2/#losses , please let me know if it is 
not correct. Thanks again.
   

----------------------------------------------------------------
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:
[email protected]


With regards,
Apache Git Services

Reply via email to