operation

Base classes for quantum operations

Classes

Channel([inputs, name, nqubit, wires, ...])	A base class for quantum channels.
Gate([name, nqubit, wires, controls, ...])	A base class for quantum gates.
GateQPD(bases, coeffs[, label, name, ...])	A base class for quasiprobability-decomposition gates.
Layer([name, nqubit, wires, den_mat, tsr_mode])	A base class for quantum layers.
MeasureQPD([nqubit, wires])	A operation for denoting a QPD measurement location.
Operation([name, nqubit, wires, den_mat, ...])	A base class for quantum operations.

class Channel(inputs: Any = None, name: str | None = None, nqubit: int = 1, wires: int | list[int] | None = None, tsr_mode: bool = False, requires_grad: bool = False)

Bases: Operation

A base class for quantum channels.

Parameters:

• inputs (Any) – The parameter of the channel. Default: None

• name (str | None) – The name of the channel. Default: None

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | None) – The indices of the qubits that the quantum operation acts on. Default: None

• tsr_mode (bool) – Whether the quantum operation is in tensor mode, which means the input and output are represented by a tensor of shape \((\text{batch}, 2, ..., 2)\). Default: False

• requires_grad (bool) – Whether the parameter is nn.Parameter or buffer. Default: False (which means buffer)

extra_repr() → str

forward(x: Tensor) → Tensor

Perform a forward pass.

get_matrix(theta: Any) → Tensor

Update the local Kraus matrices acting on density matrices.

init_para(inputs: Any = None) → None

Initialize the parameters.

inputs_to_tensor(inputs: Any = None) → Tensor

Convert inputs to torch.Tensor.

op_den_mat(x: Tensor) → Tensor

Perform a forward pass for density matrices.

property prob

The error probability.

update_matrix() → Tensor

Update the local Kraus matrices acting on density matrices.

class Gate(name: str | None = None, nqubit: int = 1, wires: int | list[int] | None = None, controls: int | list[int] | None = None, condition: bool = False, den_mat: bool = False, tsr_mode: bool = False)

Bases: Operation

A base class for quantum gates.

Parameters:

• name (str | None) – The name of the gate. Default: None

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | None) – The indices of the qubits that the quantum operation acts on. Default: None

• controls (int | list[int] | None) – The indices of the control qubits. Default: None

• condition (bool) – Whether to use controls as conditional measurement. Default: False

• den_mat (bool) – Whether the quantum operation acts on density matrices or state vectors. Default: False (which means state vectors)

• tsr_mode (bool) – Whether the quantum operation is in tensor mode, which means the input and output are represented by a tensor of shape \((\text{batch}, 2, ..., 2)\). Default: False

extra_repr() → str

forward(x: Tensor | MatrixProductState | DistributedQubitState) → Tensor | MatrixProductState | DistributedQubitState

Perform a forward pass.

get_derivative(inputs: Any) → Tensor

Get the derivative of the local unitary matrix.

get_matrix(inputs: Any) → Tensor

Get the local unitary matrix.

get_mpo() → tuple[list[Tensor], int]

Convert gate to MPO form with identities at empty sites.

Note

If sites are not adjacent, insert identities in the middle, i.e.,

>>>      |       |            |   |   |
>>>    --A---x---B--   ->   --A---I---B--
>>>      |       |            |   |   |

where

>>>         a
>>>         |
>>>    --i--I--j--
>>>         |
>>>         b

means \(\delta_{i,j} \delta_{a,b}\)

inverse() → Gate

Get the inversed gate.

op_den_mat(x: Tensor) → Tensor

Perform a forward pass for density matrices.

op_den_mat_base(x: Tensor, matrix: Tensor) → Tensor

Perform a forward pass of a gate for density matrices.

op_den_mat_control(x: Tensor, matrix: Tensor) → Tensor

Perform a forward pass of a controlled gate for density matrices.

op_dist_state(x: DistributedQubitState) → DistributedQubitState

Perform a forward pass of a gate for a distributed state vector.

op_mps(mps: MatrixProductState) → MatrixProductState

Perform a forward pass for the MatrixProductState.

op_state(x: Tensor) → Tensor

Perform a forward pass for state vectors.

op_state_base(x: Tensor, matrix: Tensor) → Tensor

Perform a forward pass of a gate for state vectors.

op_state_control(x: Tensor, matrix: Tensor) → Tensor

Perform a forward pass of a controlled gate for state vectors.

qpd(label: int | None = None) → Gate

Get the quasiprobability-decomposition representation.

set_controls(controls: int | list[int]) → None

Set the control wires of the Operation.

update_matrix() → Tensor

Update the local unitary matrix.

class GateQPD(bases: nn.ModuleList[nn.ModuleList[nn.Sequential]], coeffs: list[float], label: int | None = None, name: str | None = None, nqubit: int = 1, wires: int | list[int] | None = None, den_mat: bool = False, tsr_mode: bool = False)

Bases: Gate

A base class for quasiprobability-decomposition gates.

Parameters:

• bases (nn.ModuleList[nn.ModuleList[nn.Sequential]]) – The probabilistic basis operations for the decomposition. A nested structure where: - bases[i] is the i-th term of the QPD (corresponding to coeffs[i]). - bases[i][j] is the nn.Sequential operations applied to the j-th qubit.

• coeffs (list[float]) – The coefficients for quasiprobability representation.

• label (int | None) – The label of the gate. Default: None

• name (str | None) – The name of the quantum operation. Default: None

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | None) – The indices of the qubits that the quantum operation acts on. Default: None

• den_mat (bool) – Whether the quantum operation acts on density matrices or state vectors. Default: False (which means state vectors)

• tsr_mode (bool) – Whether the quantum operation is in tensor mode, which means the input and output are represented by a tensor of shape \((\text{batch}, 2, ..., 2)\). Default: False

forward(x: Tensor, idx: int | None = None) → Tensor

Perform a forward pass.

Parameters:

• x (Tensor) – The input tensor.

• idx (int | None) – The index of the operation to be applied. Default: None

set_nqubit(nqubit: int) → None

Set the number of qubits of the GateQPD.

set_wires(wires: int | list[int]) → None

Set the wires of the GateQPD.

class Layer(name: str | None = None, nqubit: int = 1, wires: int | list[int] | list[list[int]] | None = None, den_mat: bool = False, tsr_mode: bool = False)

Bases: Operation

A base class for quantum layers.

Parameters:

• name (str | None) – The name of the layer. Default: None

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | list[list[int]] | None) – The indices of the qubits that the quantum operation acts on. Default: None

• den_mat (bool) – Whether the quantum operation acts on density matrices or state vectors. Default: False (which means state vectors)

• tsr_mode (bool) – Whether the quantum operation is in tensor mode, which means the input and output are represented by a tensor of shape \((\text{batch}, 2, ..., 2)\). Default: False

forward(x: Tensor | MatrixProductState | DistributedQubitState) → Tensor | MatrixProductState | DistributedQubitState

Perform a forward pass.

get_unitary() → Tensor

Get the global unitary matrix.

init_para(inputs: Any = None) → None

Initialize the parameters.

inverse() → Layer

Get the inversed layer.

pattern(nodes: list[list[int]], ancilla: list[list[int]]) → Sequential

Get the MBQC pattern.

set_nqubit(nqubit: int) → None

Set the number of qubits of the Layer.

set_wires(wires: int | list[int] | list[list[int]]) → None

Set the wires of the Layer.

update_npara() → None

Update the number of parameters.

class MeasureQPD(nqubit: int = 1, wires: int | list[int] | None = None)

Bases: Operation

A operation for denoting a QPD measurement location.

Parameters:

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | None) – The indices of the qubits that the quantum operation acts on. Default: None

forward(x: Any) → Any

Perform a forward pass.

class Operation(name: str | None = None, nqubit: int = 1, wires: int | list[int] | None = None, den_mat: bool = False, tsr_mode: bool = False)

Bases: Module

A base class for quantum operations.

Parameters:

• name (str | None) – The name of the quantum operation. Default: None

• nqubit (int) – The number of qubits that the quantum operation acts on. Default: 1

• wires (int | list[int] | None) – The indices of the qubits that the quantum operation acts on. Default: None

• den_mat (bool) – Whether the quantum operation acts on density matrices or state vectors. Default: False (which means state vectors)

• tsr_mode (bool) – Whether the quantum operation is in tensor mode, which means the input and output are represented by a tensor of shape \((\text{batch}, 2, ..., 2)\). Default: False

forward(x: Tensor) → Tensor

Perform a forward pass.

get_unitary() → Tensor

Get the global unitary matrix.

init_para() → None

Initialize the parameters.

matrix_rep(x: Tensor) → Tensor

Get the density matrix representation of the state.

set_nqubit(nqubit: int) → None

Set the number of qubits of the Operation.

set_wires(wires: int | list[int]) → None

Set the wires of the Operation.

tensor_rep(x: Tensor) → Tensor

Get the tensor representation of the state.

vector_rep(x: Tensor) → Tensor

Get the vector representation of the state.