qailab.torch.orca_layer module

ORCA layer module

class qailab.torch.orca_layer.ORCALayer(in_features: int, observable: Literal['avg-photons', 'covariances', 'correlations'] = 'avg-photons', gradient_mode: Literal['parameter-shift', 'finite-difference', 'spsa'] = 'parameter-shift', gradient_delta: float = 0.3141592653589793, n_samples: int = 100, n_tiling: int = 1, tbi_type: str | None = None, n_loops: int | None = None, url: str | None = None, machine: str | None = None, secret_key: str | None = None, **tbi_params)

Bases: Module

Neural Network layer using ORCA quantum computers, a PTLayer wrapper

Parameters

in_features: int,

number fo layer input features

observable: Literal[‘avg-photons’, ‘covariances’, ‘correlations’], default = ‘avg-photons’

method of conversion from measurements to a tensor. Default is “avg-photons”.

gradient_mode: Literal[‘parameter-shift’, ‘finite-difference’, ‘spsa’], default = ‘parameter-shift’,

method to compute the gradient. Default is “parameter-shift”.

gradient_delta: float, default = np.pi / 10

Delta to use with the parameter shift rule or for the finite difference. Default is np.pi / 10.

n_samples: int, default = 100

Number of samples to draw. Default is 100.

n_tiling: int, default = 1

Uses n_tiling instances of PT Series and concatenates the results. Input features are distributed between tiles, which each have different trainable params. Default is 1.

tbi_type: Literal[‘multi-loop’, ‘single-loop’, ‘fixed-random-unitary’, ‘PT’], default = ‘single-loop’

Type of TBI to return. Can be ‘multi-loop’, ‘single-loop’, ‘fixed-random-unitary’ or ‘PT. Choose PT to run on real PT device. Default is ‘single-loop’.

n_loops: int, default = 1

Number of loops in the TBI. Default it 1.

url: str, default = None

The URL of the PT device, for example “http://<orca_api_address>”.

tbi_params: dict, default = None

Dictionary of optional parameters to instantiate the TBI. Default is None.

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

ORCA layer pytorch Module forward method

Parameters

x: torch.Tensor | None, default = None.

layer input tensor

n_samples: int | None, default = None.

number of samples from quantum computer

Returns

output: torch.Tensor

layer output tensor

set_thetas(theta_values: Tensor)

set theta parameters of layer’s beam splitter gates

Parameters

theta_values: torch.Tensor

values of theta parameter of layer’s beam splitter gates