pdft.bases.MERABasis

class pdft.bases.MERABasis(m, n, tensors=None, phases=None, code=None, inv_code=None, seed=None, parametrization='cp')

Bases: object

2D MERA basis. Each dimension with >=2 qubits must be a power of 2.

Mirror of upstream src/basis.jl:840-1070.

Parameters:

• m (int)

• n (int)

• tensors (list[Array])

• phases (Sequence[float] | None)

• code (object)

• inv_code (object)

• seed (int | None)

• parametrization (str)

__init__(m, n, tensors=None, phases=None, code=None, inv_code=None, seed=None, parametrization='cp')

When seed is provided AND phases is None, draws phases from np.random.default_rng(seed).normal(0, 0.1, n_gates) — Julia’s randn(n_gates) * 0.1 convention from ParametricDFT.jl/src/training.jl::_init_circuit.

parametrization is “cp” (diagonal disentanglers/isometries on U(1)^4) or “u4” (dense two-qubit gates on U(4), the canonical form).

Parameters:

• m (int)

• n (int)

• tensors (Sequence[Array] | None)

• phases (Sequence[float] | None)

• code (object | None)

• inv_code (object | None)

• seed (int | None)

• parametrization (str)

Methods

__init__(m, n[, tensors, phases, code, ...])	When seed is provided AND phases is None, draws phases from np.random.default_rng(seed).normal(0, 0.1, n_gates) — Julia's randn(n_gates) * 0.1 convention from ParametricDFT.jl/src/training.jl::_init_circuit.
forward_transform(pic)
inverse_transform(pic)

Attributes

image_size
inv_tensors
num_parameters
m
n
n_row_gates
n_col_gates
tensors
code
inv_code

m: int

n: int

n_row_gates: int

n_col_gates: int

tensors: list[Array]

code: object

inv_code: object

property inv_tensors: list[Array]

property image_size: tuple[int, int]

property num_parameters: int

forward_transform(pic)

Parameters:

pic (Array)

Return type:

Array

inverse_transform(pic)

Parameters:

pic (Array)

Return type:

Array