microquantum.qml.reuploading

Data-reuploading quantum classifier with a training entry point.

DataReuploadingClassifier interleaves the feature encoding with variational layers (each layer re-uploads the input features), which increases expressivity over single-uploading classifiers. It follows the VariationalClassifier predict / score contract and adds the missing fit() training loop (cross-entropy minimization through any Optimizer, with optional early-stopping callbacks).

Module Contents

class microquantum.qml.reuploading.DataReuploadingClassifier(num_features, num_classes=2, layers=2, encoder=None, optimizer=None)

Classifier with per-layer data re-uploading plus fit training.

Parameters:

• num_features (int) – Number of input features.

• num_classes (int) – Number of output classes (default 2).

• layers (int) – Number of re-uploading layers (must be >= 1).

• encoder (Optional[microquantum.qml.encoding.BaseEncoder]) – Feature map re-uploaded each layer (default AngleEncoding).

• optimizer (Optional[microquantum.optimizers.base.Optimizer]) – Classical optimizer for fit().

property num_features: int

Number of input features.

Return type:

int

property num_classes: int

Number of output classes.

Return type:

int

property layers: int

Number of re-uploading layers.

Return type:

int

property num_qubits: int

Total qubits (encoding + output).

Return type:

int

property parameters: dict[microquantum.core.parameter.Parameter, float]

Current parameter values.

Return type:

dict[microquantum.core.parameter.Parameter, float]

build_circuit(features, params)

Build the re-uploading circuit for one input.

Parameters:

• features (list[float])

• params (dict[microquantum.core.parameter.Parameter, float])

Return type:

microquantum.core.circuit.QuantumCircuit

predict(X, params=None)

Predict class labels for inputs.

Parameters:

• X (list[list[float]])

• params (Optional[dict[microquantum.core.parameter.Parameter, float]])

Return type:

microquantum.qml.classifier.ClassifierResult

score(X, y, params=None)

Classification accuracy.

Parameters:

• X (list[list[float]])

• y (list[int])

• params (Optional[dict[microquantum.core.parameter.Parameter, float]])

Return type:

float

fit(X, y, initial_params=None, callbacks=())

Train on labeled data by minimizing cross-entropy loss.

Updates the classifier’s parameters in place and returns the training-set result including the OptimizerResult.

Parameters:

• X (list[list[float]])

• y (list[int])

• initial_params (Optional[dict[microquantum.core.parameter.Parameter, float]])

• callbacks (Sequence[microquantum.optimizers.callbacks.CallbackProtocol])

Return type:

microquantum.qml.classifier.ClassifierResult