cr.sparse.dict.build_grassmannian_frame

cr.sparse.dict.build_grassmannian_frame(init, frac=0.9, shrink_mu=0.9, iterations=25)

Builds a Grassmannian frame starting from a random matrix

Parameters

• init (jax.numpy.ndarray) – initial frame

• frac (float) – Threshold for fraction of off diagonal entries to keep/change

• shrink_mu (float) – Factor by which to shrink or expand off diagonal entries

• iterations (int) – Number of iterations for alternate projections

Returns

(jax.numpy.ndarray) A frame which is approximately Grassmannian

It uses an alternate projections based algorithm.