[ONNX]Error when importing onnx model exported by pytorch

Questions
#1

Hi,

I’m trying to use pytorch to export the model with onnx format, and import that onnx folder using tvm. An error occurred at nnvm.frontend.from_onnx(onnx_model):

Traceback (most recent call last):
File “test.py”, line 29, in
sym, params = nnvm.frontend.from_onnx(onnx_model)
File “/home/ubuntu/.local/lib/python3.5/site-packages/nnvm-0.8.0-py3.5.egg/nnvm/frontend/onnx.py”, line 974, in from_onnx
sym, params = g.from_onnx(graph, opset)
File “/home/ubuntu/.local/lib/python3.5/site-packages/nnvm-0.8.0-py3.5.egg/nnvm/frontend/onnx.py”, line 829, in from_onnx
op = self._convert_operator(op_name, inputs, attr, opset)
File “/home/ubuntu/.local/lib/python3.5/site-packages/nnvm-0.8.0-py3.5.egg/nnvm/frontend/onnx.py”, line 930, in _convert_operator
sym = convert_map[op_name](inputs, attrs, self._params)
File “/home/ubuntu/.local/lib/python3.5/site-packages/nnvm-0.8.0-py3.5.egg/nnvm/frontend/onnx.py”, line 207, in _impl_v1
channels = _infer_channels(inputs[1], params, True)
IndexError: list index out of range

I’m using pytorch0.4.1, onnx 1.3.0 (installed by pip), and the master branch of dmlc/tvm on github. The model I’m testing with is Resnet50.
I tried to use caffe2.python.onnx.backend to load the onnx file that generated by pytorch, and it worked fine.

Thanks!

#2

hi Eliza~

If it is convenient, you can give me your script, onnx model, i can have a look~

#3

Hi, Hao @hao_lin ,
I also have the exact error message when importing onnx model, I wonder how can I fix this problem? Thank you~