View page source

warp.smooth_normalize#

warp.smooth_normalize(v: Any, delta: float) → Any#

  • Kernel

  • Differentiable

Normalize a vector using the pseudo-Huber norm.

See norm_pseudo_huber().

\[\frac{v}{H^\prime(v)}\]
Parameters:

  • v (Vector[Float, Any]) – The vector to normalize.

  • delta (float) – The threshold value, defaults to 1.0.

Returns:

The normalized vector.

Return type:

Vector[Float, Any]