Hi I am trying to debug during torch.onnx.export, when the layer/node is being converted, a def _parse_arg(value, desc, arg_name=None, node_name=None):
method is called (torch/onnx/symbolic_helper.py)
I tried to set a breakpoint and evaluate value.node()
in debug console (in VS Code) and it said RuntimeError: has been invalidated
Is there any way to check the nodes during debug?