ONNX symbolic expected a constant value in the trace

shridharkini6 commented 5 years ago

Describe the bug I was trying to convert torch model to onnx. I got above error from torch/onnx/symbolic.py What could be the reason

Python 2.7.12 ONNX version: 1.4.1 ONNX-TF version: 1.2.1 Tensorflow version: 1.5.0

tjingrant commented 5 years ago

We need a bit more information (specific steps to reproduce bugs, model in question if possible). From your description, I'm not even sure what you are doing is relevant to onnx-tensorflow.

I was trying to convert torch model to onnx.

Sounds like you need to post an issue with PyTorch developers.

shridharkini6 commented 5 years ago

@tjingrant please find the source code , model at https://drive.google.com/drive/folders/1QYlHnQyt0qgmBuuQR1z2B-Vn_ejzK1GA?usp=sharing

Please $git clone https://github.com/dangweili/pedestrian-attribute-recognition-pytorch.git $mv peta_dataset.pkl pedestrian-attribute-recognition-pytorch/dataset/peta/

Please let me know of you find any issues running it Python 2.7.12 ONNX version: 1.4.1 ONNX-TF version: 1.2.1 Tensorflow version: 1.5.0

fumihwh commented 5 years ago

@shridharkini6 Could you paste your error log here? Not just let us to run it.

lookup1980 commented 5 years ago

I had a same issue, and the error log is:

Traceback (most recent call last):
  File "converter.py", line 178, in <module>
    torch.onnx.export(model, img, "model.onnx", verbose=True)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/__init__.py", line 27, in export
    return utils.export(*args, **kwargs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 104, in export
    operator_export_type=operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 281, in _export
    example_outputs, propagate)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 227, in _model_to_graph
    graph = _optimize_graph(graph, operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 155, in _optimize_graph
    graph = torch._C._jit_pass_onnx(graph, operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/__init__.py", line 52, in _run_symbolic_function
    return utils._run_symbolic_function(*args, **kwargs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 504, in _run_symbolic_function
    return fn(g, *inputs, **attrs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 88, in wrapper
    args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 88, in <listcomp>
    args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 45, in _parse_arg
    raise RuntimeError("ONNX symbolic expected a constant value in the trace")
RuntimeError: ONNX symbolic expected a constant value in the trace

vinamrasb commented 5 years ago

I am also facing the same problem.

chinhuang007 commented 5 years ago

The error seems coming from the pytorch converter, not onnx-tf. Please clarify and provide exact code that invokes onnx-tf backend converter. We will close the issue if determined not related to onnx-tf.

clw5180 commented 5 years ago

I had a same issue, and the error log is:

Traceback (most recent call last):
  File "converter.py", line 178, in <module>
    torch.onnx.export(model, img, "model.onnx", verbose=True)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/__init__.py", line 27, in export
    return utils.export(*args, **kwargs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 104, in export
    operator_export_type=operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 281, in _export
    example_outputs, propagate)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 227, in _model_to_graph
    graph = _optimize_graph(graph, operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 155, in _optimize_graph
    graph = torch._C._jit_pass_onnx(graph, operator_export_type)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/__init__.py", line 52, in _run_symbolic_function
    return utils._run_symbolic_function(*args, **kwargs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/utils.py", line 504, in _run_symbolic_function
    return fn(g, *inputs, **attrs)
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 88, in wrapper
    args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 88, in <listcomp>
    args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
  File "/home/mgu/anaconda3/envs/converter-gpu36/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 45, in _parse_arg
    raise RuntimeError("ONNX symbolic expected a constant value in the trace")
RuntimeError: ONNX symbolic expected a constant value in the trace

same error. Have you fixed that?

onnx / onnx-tensorflow

ONNX symbolic expected a constant value in the trace #360