lucid.zeros_like

lucid.zeros_like(a: Tensor | list | ndarray | array, /, dtype: type = None, requires_grad: bool = False, keep_grad: bool = False, device: Literal['cpu', 'gpu'] | None = None) → Tensor

The zeros_like function creates a tensor of zeros with the same shape and optionally the same data type as a given input tensor.

Function Signature

def zeros_like(
    a: Tensor | _ArrayLike,
    dtype: Any = None,
    requires_grad: bool = False,
    keep_grad: bool = False,
    device: _DeviceType | None = None,
) -> Tensor

Parameters

• a (Tensor | _ArrayLike):

  The input tensor or array whose shape will be used to create the zero-filled tensor.

• dtype (Any, optional):

  The data type of the elements in the tensor. If None, the data type of a will be used. Defaults to None.

• requires_grad (bool, optional):

  If True, the resulting tensor will be part of the computation graph and capable of tracking gradients. Defaults to False.

• keep_grad (bool, optional):

  If True, the gradient history will be preserved even if the tensor does not require gradients. Defaults to False.

Returns

• Tensor:

  A tensor filled with zeros, having the same shape as the input tensor a and optionally the same data type.

Example

>>> import lucid
>>> a = Tensor([[1, 2, 3], [4, 5, 6]])
>>> z = lucid.zeros_like(a)
>>> print(z)
Tensor([[0. 0. 0.]
        [0. 0. 0.]])

You can specify a different data type or gradient settings:

>>> z = lucid.zeros_like(a, dtype=np.float64, requires_grad=True)
>>> print(z.dtype, z.requires_grad)
float64 True

Note

• This function is useful for creating zero-initialized tensors that mirror the shape and optionally the type of another tensor or array.

• If dtype is None, the data type of a is automatically used for the output tensor.

• The requires_grad parameter is useful for tensors involved in gradient-based computation graphs.

• The keep_grad parameter allows preserving gradient history even for tensors that do not require gradients.