xma.functional.p_norm

p_norm(x: Tensor, multiplier: float | None = None, p: int | str = 2, output_dtype: dtype = torch.float32, *, kernel_backend: KernelBackend | None = None) → Tensor

computes norm of a vector

Parameters:

• x (torch.Tensor) – input activation

• multiplier (float | None) – if not None, pre-multiplies x with multiplier. Defaults to None.

• p (int | str) – norm type. can be integer >= 1 or inf

• output_dtype (torch.dtype) – output dtype

• kernel_backend (KernelBackend | None) – KernelBackend

Returns:

output activation

Return type:

Tensor