t3_linalg

Linear algebra on ragged t3 data, defined w.r.t. the REPRESENTED dense tensor.

t3_add/t3_scale/t3_inner_product/t3_norm/t3_mult and the t3m_* forms simulate operations on the dense tensor, using the cores only as a computational device (add concatenates ranks; nothing here is corewise). t3_sum_stack is the rank-growing tensor sum over a stack – not the rank-preserving corewise core sum.

Functions

t3_add(x, y)	Add Tucker tensor trains x and y, yielding a Tucker tensor train with summed ranks.
t3_sum_stack(x[, axis])	Sum the dense tensors represented by a stacked Tucker tensor train over stack axes.
t3_scale(x, s)	Multipy a Tucker tensor train by a scaling factor.
t3_inner_product(x, y[, use_orthogonalization])	Compute Hilbert-Schmidt inner product of two Tucker tensor trains.
t3_norm(x[, use_orthogonalization])	Compute Hilbert-Schmidt norm of a Tucker tensor train.
t3_mult(x, y)	Pointwise multiply Tucker tensor trains x and y, yielding a Tucker tensor train with multiplied ranks.
t3m_form_then_round(x, y[, max_tucker_ranks, ...])	Elementwise product x ⊙ y -- method (a): form the full product, then round.
t3m_inplace_fused(x, y[, max_tucker_ranks, ...])	Elementwise product x ⊙ y -- method (b): a fused left-to-right sweep that truncates as it
t3m_swap(x, y[, max_tucker_ranks, max_tt_ranks, rtol, ...])	Elementwise product x ⊙ y -- method (c): the swap-based TTM generalization, for
t3_plus_scalar(x, s)	Add a scalar to a Tucker tensor train: (x + s).to_dense() == x.to_dense() + s.
t3_weighted_norm(x0, weights[, use_orthogonalization])	Weighted Hilbert-Schmidt norm of a Tucker tensor train: t3_norm(absorb(x0, weights)) -- the norm
t3_weighted_inner(x0_A, weights_A, x0_B, weights_B[, ...])	Weighted Hilbert-Schmidt inner product <absorb(A), absorb(B)> of two weighted Tucker tensor