Function tensor

python/paddle/tensor/creation.py:980–1099

r""" Constructs a ``paddle.Tensor`` from ``data`` , which can be scalar, tuple, list, numpy\.ndarray, paddle\.Tensor. If the ``data`` is already a Tensor, copy will be performed and return a new tensor. If you only want to change stop_gradient property, please call ``Tensor.stop_gra

(
    data: TensorLike | NestedNumericSequence,
    dtype: DTypeLike | None = None,
    device: PlaceLike | None = None,
    requires_grad: bool = False,
    pin_memory: bool = False,
)

def tensor(
    data: TensorLike | NestedNumericSequence,
    dtype: DTypeLike | None = None,
    device: PlaceLike | None = None,
    requires_grad: bool = False,
    pin_memory: bool = False,
) -> paddle.Tensor:
    r"""
    Constructs a ``paddle.Tensor`` from ``data`` ,
    which can be scalar, tuple, list, numpy\.ndarray, paddle\.Tensor.

    If the ``data`` is already a Tensor, copy will be performed and return a new tensor.
    If you only want to change stop_gradient property, please call ``Tensor.stop_gradient = stop_gradient`` directly.

    .. code-block:: text

        We use the dtype conversion rules following this:
                Keep dtype
        np.number ───────────► paddle.Tensor
                                (0-D Tensor)
                    default_dtype
        Python Number ───────────────► paddle.Tensor
                                        (0-D Tensor)
                    Keep dtype
        np.ndarray ───────────► paddle.Tensor

    Args:
        data(scalar|tuple|list|ndarray|Tensor): Initial data for the tensor.
            Can be a scalar, list, tuple, numpy\.ndarray, paddle\.Tensor.
        dtype(str|np.dtype, optional): The desired data type of returned tensor. Can be 'bool' , 'float16' ,
            'float32' , 'float64' , 'int8' , 'int16' , 'int32' , 'int64' , 'uint8',
            'complex64' , 'complex128'. Default: None, infers dtype from ``data``
            except for python float number which gets dtype from ``get_default_type`` .
        device(CPUPlace|CUDAPinnedPlace|CUDAPlace|str, optional): The place to allocate Tensor. Can be
            CPUPlace, CUDAPinnedPlace, CUDAPlace. Default: None, means global place. If ``device`` is
            string, It can be ``cpu``, ``gpu:x`` and ``gpu_pinned``, where ``x`` is the index of the GPUs.
        requires_grad(bool, optional): Whether to block the gradient propagation of Autograd. Default: False.
        pin_memory(bool, optional): If set, return tensor would be allocated in the pinned memory. Works only for CPU tensors. Default: False

    Returns:
        Tensor: A Tensor constructed from ``data`` .

    Examples:
        .. code-block:: pycon

            >>> # type: ignore
            >>> import paddle

            >>> type(paddle.tensor(1))
            <class 'paddle.Tensor'>

            >>> paddle.tensor(1)
            Tensor(shape=[], dtype=int64, place=Place(cpu), stop_gradient=True,
            1)

            >>> x = paddle.tensor(1, requires_grad=True)
            >>> print(x)
            Tensor(shape=[], dtype=int64, place=Place(cpu), stop_gradient=False,