normalize

normalize(data: Tensor, order: int | float, scale: float | complex = 1.0, dim: int | tuple[int, ...] | None = None, epsilon: float = 1e-12) → Tensor

Normalizes a tensor.

Parameters:

• data (torch.Tensor) – data to normalize.

• order (int | float) – order of \(p\)-norm by which to normalize.

• scale (float | complex, optional) – desired \(p\)-norm of elements along specified dimensions. Defaults to 1.0.

• dim (int | tuple[int, ...] | None, optional) – dimension(s) along which to normalize, all dimensions if None. Defaults to None.

• epsilon (float, optional) – value added to the denominator in case of zero-valued norms. Defaults to 1e-12.

Returns:

normalized tensor.

Return type:

torch.Tensor