Minimal setup for ONNX incongruence

import torch


class DummyModel(torch.nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, d1, d2, d3):
        shape = [d1, d2, d3]
        """
        Option 1: convert through tensor.long()
        No warning
        """
        shape = [dim.long() for dim in shape]
        """
        Option 2: convert through torch.tensor(*, dtype=torch.int64)
        TracerWarning: torch.tensor results are registered as constants in the trace.
        """
        # shape = [torch.tensor(dim, dtype=torch.int64) for dim in shape]
        """
        Option 3: don't convert, keep int32
        UserWarning: The exported ONNX model failed ONNX shape inference.
        """
        # <no code>
        return torch.full(shape, 2)


dummy_model = DummyModel()
d1 = torch.tensor(2, dtype=torch.int32)
d2 = torch.tensor(2, dtype=torch.int32)
d3 = torch.tensor(2, dtype=torch.int32)
torch.onnx.export(
    dummy_model,
    (d1, d2, d3),
    f="my_filename.onnx",
    verbose=True,
    input_names=["d1", "d2", "d3"],
    output_names=["casted_data"],
)