View Source Scholar.Preprocessing.Normalizer (Scholar v0.3.0)

Implements functionality for rescaling tensor to unit norm. It enables to apply normalization along any combination of axes.

Summary

Functions

fit_transform(tensor, opts \\ [])

Normalize samples individually to unit norm.

Functions

Link to this function

fit_transform(tensor, opts \\ [])

View Source

Normalize samples individually to unit norm.

The zero-tensors cannot be normalized and they stay the same after normalization.

Options

  • :axes - Axes to calculate the distance over. By default the distance is calculated between the whole tensors.

  • :norm - The norm to use to normalize each non zero sample. Possible options are :euclidean, :manhattan, and :chebyshev The default value is :euclidean.

Examples 

iex> t = Nx.tensor([[0, 0, 0], [3, 4, 5], [-2, 4, 3]])
iex> Scholar.Preprocessing.Normalizer.fit_transform(t, axes: [1])
#Nx.Tensor<
  f32[3][3]
  [
    [0.0, 0.0, 0.0],
    [0.4242640733718872, 0.5656854510307312, 0.7071067690849304],
    [-0.3713906705379486, 0.7427813410758972, 0.5570860505104065]
  ]
>

iex> t = Nx.tensor([[0, 0, 0], [3, 4, 5], [-2, 4, 3]])
iex> Scholar.Preprocessing.Normalizer.fit_transform(t)
#Nx.Tensor<
  f32[3][3]
  [
    [0.0, 0.0, 0.0],
    [0.3375263810157776, 0.4500351846218109, 0.5625439882278442],
    [-0.22501759231090546, 0.4500351846218109, 0.3375263810157776]
  ]
>