These warnings from Einops package .../einops/einops.py are found during conversion:
TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
unknown: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] == _unknown_axis_length}
However, no exceptions raised from this code. I am not sure whether the tracer warnings caused by Einops are related to the indefinite run of torch.onnx.export().
When converting Torch model to ONNX, the conversion ran for more than 8 hours without any exception.
These warnings from Einops package .../einops/einops.py are found during conversion:
However, no exceptions raised from this code. I am not sure whether the tracer warnings caused by Einops are related to the indefinite run of torch.onnx.export().
I appreciate any help for this issue, thank you.