torch_xla scan forces inputs to have gradients

[tengyifei]

The snippet

xla/torch_xla/experimental/scan.py

Lines 217 to 226 in 00fac78

	# Make some fake tensors to trace the user function and obtain the
	# forward and backward graphs. Note that the init/carry fake tensor
	# always requires grad. That's because even if the user passed in some
	# `init` that does not require grad, we still want gradients to flow
	# through the `carry` from one iteration of the user function to the
	# next. In summary, the `carry` argument used to trace a user function
	# to get a correct backward pass always requires grad.
	def make_fake_tensor(v: torch.Tensor, requires_grad=True) -> torch.Tensor:
	return torch.empty_like(
	v, dtype=v.dtype, device=v.device, requires_grad=requires_grad)

is probably wrong.

The most obvious example is that if one of the input is an integer, then it can't possibly have gradients.