Module vis: visualization methods for tensors

Package teneva, module vis: visualization methods for tensors.

This module contains the functions for visualization of TT-tensors.

---

teneva_jax.vis.show(Y)

Check and display mode size and TT-rank of the given TT-tensor.

Parameters:

Y (list) – TT-tensor.

Examples:

# 5-dim random TT-tensor with mode size 4 and TT-rank 3:
rng, key = jax.random.split(rng)
Y = teneva.rand(5, 4, 3, key)

# Print the resulting TT-tensor:
teneva.show(Y)

# >>> ----------------------------------------
# >>> Output:

# TT-tensor-jax | d =     5 | n =     4 | r =     3 |
#

If an incorrect TT-tensor is passed to the function (the correctness of the shape of all cores is explicitly checked), then an error will be generated:

Y = []

try:
    teneva.show(Y)
except ValueError as e:
    print('Error :', e)

# >>> ----------------------------------------
# >>> Output:

# Error : Invalid TT-tensor
#

Y = [42.]

try:
    teneva.show(Y)
except ValueError as e:
    print('Error :', e)

# >>> ----------------------------------------
# >>> Output:

# Error : Invalid TT-tensor
#

Y = [
    jnp.zeros((1, 5, 7)),
    jnp.zeros((100, 42, 7, 1)),
    jnp.zeros((42, 7, 1))]

try:
    teneva.show(Y)
except ValueError as e:
    print('Error :', e)

# >>> ----------------------------------------
# >>> Output:

# Error : Invalid shape of middle cores for TT-tensor
#

import numpy as onp # Numpy is not supported!

Y = [
    onp.zeros((1, 5, 3)),
    onp.zeros((100, 3, 5, 3)),
    onp.zeros((3, 5, 1))]

try:
    teneva.show(Y)
except ValueError as e:
    print('Error :', e)

# >>> ----------------------------------------
# >>> Output:

# Error : Invalid left core of TT-tensor
#

rng, key = jax.random.split(rng)
Y = teneva.rand(5, 6, 7, key)

try:
    teneva.show(Y)
except ValueError as e:
    print('Error :', e)

# >>> ----------------------------------------
# >>> Output:

# Error : TT-rank should be no greater than mode size
#