You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2018/05/09 21:29:13 UTC

[GitHub] altosaar opened a new issue #10875: How to debug hybridize() failures?

altosaar opened a new issue #10875: How to debug hybridize() failures?
URL: https://github.com/apache/incubator-mxnet/issues/10875
 
 
   My model runs fine without hybridize, but I need the speed boost.
   
   When I call hybridize I get a cryptic failure message:
   
   ```Traceback (most recent call last):
     File "normal_def/train.py", line 225, in <module>
       train()
     File "normal_def/train.py", line 178, in train
       users, items, item_counts, set_sizes, user_meals)
     File "/usr/local/anaconda3/lib/python3.6/site-packages/mxnet/gluon/block.py", line 413, in __call__
       return self.forward(*args)
     File "/usr/local/anaconda3/lib/python3.6/site-packages/mxnet/gluon/block.py", line 621, in forward
       return self._call_cached_op(x, *args)
     File "/usr/local/anaconda3/lib/python3.6/site-packages/mxnet/gluon/block.py", line 528, in _call_cached_op
       out = self._cached_op(*cargs)
     File "/usr/local/anaconda3/lib/python3.6/site-packages/mxnet/_ctypes/ndarray.py", line 149, in __call__
       ctypes.byref(out_stypes)))
     File "/usr/local/anaconda3/lib/python3.6/site-packages/mxnet/base.py", line 149, in check_call
       raise MXNetError(py_str(_LIB.MXGetLastError()))
   mxnet.base.MXNetError: Error in operator normaldef0_fastpoissonlogprob0__mul0: [17:27:23] src/operator/nn/../tensor/../elemwise_op_common.h:123: Check failed: assign(&dattr, (*vec)[i]) Incompatible attr in node normaldef0_fastpoissonlogprob0__mul0 at 1-th input: expected [9963,128], got [9963,1]
   ```
   
   There are many multiplies and nodes of that shape. How can I figure out why hybridize is not working? i.e. what line of code `normaldef0_fastpoissonlogprob0__mul0` corresponds to?
   
   Thanks!

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