deel.lip.normalizers module

This module contains computation function, for BjorckNormalizer and spectral normalization. This is done for internal use only.

deel.lip.normalizers.bjorck_normalization(w, niter=15)

apply Bjorck normalization on w.

Parameters

• w – weight to normalize, in order to work properly, we must have max_eigenval(w) ~= 1

• niter – number of iterations

Returns

the orthonormal weights

deel.lip.normalizers.spectral_normalization(kernel, u=None, niter=3)

Normalize the kernel to have it’s max eigenvalue == 1.

Parameters

• kernel – the kernel to normalize

• u – initialization for the max eigen vector

• niter – number of iteration

Returns

the normalized kernel w_bar, it’s shape, the maximum eigen vector, and the maximum eigen value