ut3svd_orthogonal_representations

t3toolbox.uniform_frame_variations_format.ut3svd_orthogonal_representations(x, **t3svd_kwargs)

def ut3svd_orthogonal_representations(
        x: ut3.UniformTuckerTensorTrain,
        **t3svd_kwargs,                 # passed to UniformTuckerTensorTrain.t3svd (max_*_ranks, sharing, ...)
) -> typ.Tuple[
    UT3Frame,            # orthogonal frame at the t3svd result, in the t3svd GAUGE
    UT3Variations,       # the variations of that representation
    'ut3.UT3Weights',    # the singular values (the result's masks), ready for UT3FrameWeights.from_ut3weights
]:

Uniform twin of t3svd_orthogonal_representations(): the orthogonal frame of x in the T3-SVD gauge (already_left_orthogonal=True, so the Tucker basis is the singular basis and the returned singular values weight the right coordinates), with one SVD instead of two.

Parameters:

x (UniformTuckerTensorTrain)

Return type:

t3toolbox.backend.common.typ.Tuple[UT3Frame, UT3Variations, UT3Weights]