t3toolbox.uniform_tucker_tensor_train.ut3_to_t3

t3toolbox.uniform_tucker_tensor_train.ut3_to_t3(x_uniform: UniformTuckerTensorTrain, stack_t3s: bool = False, use_jax: bool = False) → t3toolbox.tucker_tensor_train.TuckerTensorTrain

Convert UniformTuckerTensorTrain to TuckerTensorTrain.

Examples

>>> import numpy as np
>>> import t3toolbox.tucker_tensor_train as t3
>>> import t3toolbox.uniform_tucker_tensor_train as ut3
>>> x = t3.t3_corewise_randn((14,15,16), (4,6,5), (1,3,2,1), stack_shape=(2,))
>>> uniform_x = ut3.t3_to_ut3(x) # Convert t3 -> ut3
>>> print(uniform_x.uniform_structure)
(3, 16, 6, 3, (2,))
>>> print(uniform_x.shape)
[[14 14]
 [15 15]
 [16 16]]
>>> print(uniform_x.tucker_ranks)
[[4 4]
 [6 6]
 [5 5]]
>>> print(uniform_x.tt_ranks)
[[1 1]
 [3 3]
 [2 2]
 [1 1]]
>>> all_x2 = ut3.ut3_to_t3(uniform_x) # Convert ut3 -> t3 without stacking
>>> for x2i in all_x2: print(x2i.uniform_structure)
((14, 15, 16), (4, 6, 5), (1, 3, 2, 1), ())
((14, 15, 16), (4, 6, 5), (1, 3, 2, 1), ())
>>> stacked_x2 = ut3.ut3_to_t3(uniform_x, stack_t3s=True) # with stacking
>>> print(stacked_x2.uniform_structure)
((14, 15, 16), (4, 6, 5), (1, 3, 2, 1), (2,))
>>> for B, B2 in zip(stacked_x2.tucker_cores, x.tucker_cores): print(np.linalg.norm(B - B2))
0.0
0.0
0.0
>>> for G, G2 in zip(stacked_x2.tt_cores, x.tt_cores): print(np.linalg.norm(G - G2))
0.0
0.0
0.0