PyTorch > TorchScript tensor.view() alternative?

deployment
1

0

I have the following code:

@torch.jit.script
def my_function(t: torch.Tensor) -> torch.Tensor:
    if t.dtype != torch.bfloat16:
        raise ValueError("Input tensor must be of dtype torch.bfloat16")
    int_type = torch.int16
    t_bits = t.view(int_type) <- error

That throws:

RuntimeError: shape '[2]' is invalid for input of size 1

Because, given the documentation:

.. warning::
This overload is not supported by TorchScript, and using it in a 
Torchscript program will cause undefined behavior.

So the question is, what’s the alternative for use?