torch.index_reduce

torch.index_reduce(input: Tensor, dim: int, index: Tensor, source: Tensor, reduce: str, *, include_self: bool = True, out: Optional[Tensor]) → Tensor

See index_reduce_() for function description.