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 2019/03/10 13:47:08 UTC

[GitHub] [incubator-mxnet] turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx

turtleizzy edited a comment on issue #13949: Error: shape inconsistent while converting PyTorch model to mxnet model with onnx
URL: https://github.com/apache/incubator-mxnet/issues/13949#issuecomment-471307657
 
 
   > @wangliye00 @Con-Mi However, I do see the error that you facing with MXNet v1.3.1. For fixing this, could you try pulling in the commit #13413 and checking if you are able to proceed further?
   
   I am facing similar issue when loading pytorch-densenet onnx model into mxnet. The error message reads:
   
   ```/usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/import_onnx.py in _convert_operator(self, node_name, op_name, attrs, inputs)
        59         """
        60         if op_name in convert_map:
   ---> 61             op_name, new_attrs, inputs = convert_map[op_name](attrs, inputs, self)
        62         else:
        63             raise NotImplementedError("Operator {} not implemented.".format(op_name))
   
   /usr/local/lib/python3.6/site-packages/mxnet/contrib/onnx/onnx2mx/_op_translations.py in reshape(attrs, inputs, proto_obj)
       432     if len(inputs) == 1:
       433         return 'reshape', attrs, inputs[0]
   --> 434     reshape_shape = list(proto_obj._params[inputs[1].name].asnumpy())
       435     reshape_shape = [int(i) for i in reshape_shape]
       436     new_attrs = {'shape': reshape_shape}
   
   KeyError: 'concat51'```
   
    I tried mxnet 1.3.1 (after patched `import_onnx.py` following your suggestion) and 1.4.0 with no luck, both raised similar Exception.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services