QubitUnitaryFast#

class torchquantum.operator.QubitUnitaryFast(has_params: bool = False, trainable: bool = False, init_params=None, n_wires=None, wires=None)[source]#

Bases: Operation

Class for fast implementation of controlled Qubit Unitary gate.

__init__(has_params: bool = False, trainable: bool = False, init_params=None, n_wires=None, wires=None)[source]#

_summary_

Parameters:

  • has_params (bool, optional) – Whether the operations has parameters. Defaults to False.

  • trainable (bool, optional) – Whether the parameters are trainable (if contains parameters). Defaults to False.

  • init_params (torch.Tensor, optional) – Initial parameters. Defaults to None.

  • n_wires (int, optional) – Number of qubits. Defaults to None.

  • wires (Union[int, List[int]], optional) – Which qubit the operation is applied to. Defaults to None.

Methods

build_params(trainable)[source]#

Build parameters.

Parameters:

trainable (bool) – Whether the parameters are trainable.

Returns:

Built parameters.

Return type:

torch.Tensor

classmethod from_controlled_operation(op, c_wires, t_wires, trainable)[source]#
Parameters:

  • op – the operation

  • c_wires – controlled wires, will only be a set such as 1, [2,3]

  • t_wires – can be a list of list of wires, multiple sets

  • [[1 –

  • 2] –

  • [3 –

  • 4]] –

  • trainable –

static func(q_device, wires, params=None, n_wires=None, static=False, parent_graph=None, inverse=False, comp_method='bmm')#

Perform the qubitunitaryfast gate.

Parameters:

  • q_device (tq.QuantumDevice) – The QuantumDevice.

  • wires (Union[List[int], int]) – Which qubit(s) to apply the gate.

  • params (torch.Tensor, optional) – Parameters (if any) of the gate. Default to None.

  • n_wires (int, optional) – Number of qubits the gate is applied to. Default to None.

  • static (bool, optional) – Whether use static mode computation. Default to False.

  • parent_graph (tq.QuantumGraph, optional) – Parent QuantumGraph of current operation. Default to None.

  • inverse (bool, optional) – Whether inverse the gate. Default to False.

  • comp_method (bool, optional) – Use ‘bmm’ or ‘einsum’ method to perform

  • 'bmm'. (matrix vector multiplication. Default to) –

Returns:

None.

reset_params(init_params=None)[source]#

Reset parameters.

Parameters:

init_params (torch.Tensor, optional) – Input the initialization parameters. Defaults to None.

Attributes

num_params = -1#
num_wires = -1#
op_name = 'qubitunitaryfast'#
training: bool#