gram_schmidt

tenpy.linalg.lanczos.gram_schmidt(vecs, rcond=1e-14, verbose=0)[source]

In place Gram-Schmidt Orthogonalization and normalization for npc Arrays.

Parameters
  • vecs (list of Array) – The vectors which should be orthogonalized. All with the same order of the legs. Entries are modified in place. if a norm < rcond, the entry is set to None.

  • rcond (float) – Vectors of norm < rcond (after projecting out previous vectors) are discarded.

  • verbose (int) – Print additional output if verbose >= 1.

Returns

  • vecs (list of Array) – The ortho-normalized vectors (without any None).

  • ov (2D Array) – For j >= i, ov[j, i] = npc.inner(vecs[j], vecs[i], 'range', do_conj=True) (where vecs[j] was orthogonalized to all vecs[k], k < i).