qxmt.experiment.executor module

qxmt.experiment.executor module#

class qxmt.experiment.executor.QKernelExecutor(experiment)

Bases: RunExecutorBase

Executor for QKernel models (classification / regression).

This class handles the execution of quantum kernel-based machine learning models, including both classification and regression tasks.

Parameters:

experiment (Experiment)

run_from_config(*, config, commit_id, run_dirc, n_jobs, show_progress, repo_path, add_results=True)

Run a QKernel model using configuration settings.

Parameters:

  • config (ExperimentConfig) – Configuration settings for the experiment.

  • commit_id (str) – Git commit ID for the current run.

  • run_dirc (Path) – Directory to store run artifacts.

  • n_jobs (int) – Number of parallel jobs to run.

  • show_progress (bool) – Whether to display progress bars.

  • repo_path (Optional[str]) – Path to the repository.

  • add_results (bool, optional) – Whether to save results. Defaults to True.

Returns:

Tuple containing the run artifact and record.

Return type:

tuple[RunArtifact, RunRecord]

Note

This method builds the dataset and model from the configuration before executing the run.

run_from_instance(*, task_type, dataset, model, save_shots_path, save_model_path, default_metrics_name, custom_metrics, desc, commit_id, config_file_name, repo_path, add_results=True)

Run a QKernel model using pre-built instances.

This method executes the model training, validation, and testing phases, and records the results.

Parameters:

  • task_type (Optional[str]) – Type of the task (classification/regression).

  • dataset (Dataset) – Dataset to use for training and evaluation.

  • model (BaseMLModel) – Pre-built model instance.

  • save_shots_path (Optional[Path]) – Path to save shot results.

  • save_model_path (Path) – Path to save the trained model.

  • default_metrics_name (Optional[list[str]]) – List of default metrics to use.

  • custom_metrics (Optional[list[dict[str, Any]]]) – List of custom metrics.

  • desc (str) – Description of the run.

  • commit_id (str) – Git commit ID.

  • config_file_name (Path) – Name of the configuration file.

  • repo_path (Optional[str]) – Path to the repository.

  • add_results (bool, optional) – Whether to save results. Defaults to True.

Returns:

Tuple containing the run artifact and record.

Return type:

tuple[RunArtifact, RunRecord]

class qxmt.experiment.executor.RunExecutorBase(experiment)

Bases: ABC

Abstract base class for experiment run executors.

This class provides common functionality and attributes for concrete executor implementations. It serves as a base class for both QKernel and VQE executors.

Parameters:

experiment (Experiment)

exp

Reference to the experiment instance.

Type:

Experiment

logger

Logger instance for logging operations.

eval_factory

Factory for creating evaluation instances.

Type:

EvaluationFactory

__init__(experiment)

Initialize the base executor.

Parameters:

experiment (Experiment) – The experiment instance to use for execution.

Return type:

None

abstractmethod run_from_config(*args, **kwargs)

Run a model using configuration settings.

Parameters:

  • args (Any)

  • kwargs (Any)

Return type:

tuple[RunArtifact, RunRecord]

abstractmethod run_from_instance(*args, **kwargs)

Run a model using a pre-built instance.

Parameters:

  • args (Any)

  • kwargs (Any)

Return type:

tuple[RunArtifact, RunRecord]

class qxmt.experiment.executor.VQEExecutor(experiment)

Bases: RunExecutorBase

Executor for Variational Quantum Eigensolver models.

This class handles the execution of VQE models, which are used for finding the ground state energy of quantum systems.

Parameters:

experiment (Experiment)

run_from_config(*, config, commit_id, run_dirc, n_jobs, show_progress, repo_path, add_results=True)

Run a VQE model using configuration settings.

Parameters:

  • config (ExperimentConfig) – Configuration settings for the experiment.

  • commit_id (str) – Git commit ID for the current run.

  • run_dirc (Path) – Directory to store run artifacts.

  • n_jobs (int) – Number of parallel jobs to run.

  • show_progress (bool) – Whether to display progress bars.

  • repo_path (Optional[str]) – Path to the repository.

  • add_results (bool, optional) – Whether to save results. Defaults to True.

Returns:

Tuple containing the run artifact and record.

Return type:

tuple[RunArtifact, RunRecord]

run_from_instance(*, model, save_shots_path, default_metrics_name, custom_metrics, desc, commit_id, config_file_name, repo_path, add_results=True)

Run a VQE model using a pre-built instance.

This method executes the VQE optimization process and records the results.

Parameters:

  • model (BaseVQE) – Pre-built VQE model instance.

  • save_shots_path (Optional[Path]) – Path to save shot results.

  • default_metrics_name (Optional[list[str]]) – List of default metrics to use.

  • custom_metrics (Optional[list[dict[str, Any]]]) – List of custom metrics.

  • desc (str) – Description of the run.

  • commit_id (str) – Git commit ID.

  • config_file_name (Path) – Name of the configuration file.

  • repo_path (Optional[str]) – Path to the repository.

  • add_results (bool, optional) – Whether to save results. Defaults to True.

Returns:

Tuple containing the run artifact and record.

Return type:

tuple[RunArtifact, RunRecord]

Raises:

NotImplementedError – If remote machine execution is attempted.