Task Plugins Collection

Note

See the Glossary for the meaning of the acronyms used in this guide.

Artifacts

mlflow.download_all_artifacts_in_run(run_id, ...)	Downloads an artifact file or directory from a previous MLFlow run.
mlflow.upload_data_frame_artifact(...[, ...])	Uploads a DataFrame as an artifact of the active MLFlow run.
mlflow.upload_directory_as_tarball_artifact(...)	Archives a directory and uploads it as an artifact of the active MLFlow run.
mlflow.upload_file_as_artifact(artifact_path)	Uploads a file as an artifact of the active MLFlow run.
utils.extract_tarfile(filepath[, ...])	Extracts a tarball archive into the current working directory.
utils.make_directories(dirs)	Creates directories if they do not exist.

Exceptions

exceptions.UnsupportedDataFrameFileFormatError

The requested data frame file format is not supported.

Attacks

fgm.create_adversarial_fgm_dataset(data_dir, ...)

Generates an adversarial dataset using the Fast Gradient Method attack.

Backend Configuration

tensorflow.init_tensorflow(seed)

Initializes Tensorflow to ensure compatibility and reproducibility.

Data

tensorflow.create_image_dataset(data_dir, ...)	Yields an iterator for generating batches of real-time augmented image data.
tensorflow.get_n_classes_from_directory_iterator(ds)	Returns the number of unique labels found by the tf.keras.preprocessing.image.DirectoryIterator.

Estimators

keras_classifiers.init_classifier(...[, loss])	Initializes an untrained neural network image classifier for Tensorflow/Keras.
methods.fit(estimator[, x, y, fit_kwargs])	Fits the estimator to the given data.
methods.predict(estimator[, x, predict_kwargs])	Uses the estimator to make predictions on the given input data.

Available Estimators

keras_classifiers.shallow_net(input_shape, ...)	Builds an untrained shallow neural network architecture for Tensorflow/Keras.
keras_classifiers.le_net(input_shape, n_classes)	Builds an untrained LeNet-5 neural network architecture for Tensorflow/Keras.
keras_classifiers.alex_net(input_shape, ...)	Builds an untrained AlexNet neural network architecture for Tensorflow/Keras.

Metrics

distance.get_distance_metric_list(request)	Gets multiple distance metric functions from the registry.
distance.get_distance_metric(func)	Gets a distance metric function from the registry.
performance.get_performance_metric_list(request)	Gets multiple performance metric functions from the registry.
performance.get_performance_metric(func)	Gets a performance metric function from the registry.

Available Metrics

distance.l_inf_norm(y_true, y_pred)	Calculates the L∞ norm between a batch of two matrices.
distance.l_1_norm(y_true, y_pred)	Calculates the L1 norm between a batch of two matrices.
distance.l_2_norm(y_true, y_pred)	Calculates the L2 norm between a batch of two matrices.
distance.paired_cosine_similarities(y_true, ...)	Calculates the cosine similarity between a batch of two matrices.
distance.paired_euclidean_distances(y_true, ...)	Calculates the Euclidean distance between a batch of two matrices.
distance.paired_manhattan_distances(y_true, ...)	Calculates the Manhattan distance between a batch of two matrices.
distance.paired_wasserstein_distances(...)	Calculates the Wasserstein distance between a batch of two matrices.
performance.accuracy(y_true, y_pred, **kwargs)	Calculates the accuracy score.
performance.roc_auc(y_true, y_pred, **kwargs)	Calculates the Area Under the Receiver Operating Characteristic Curve (ROC AUC).
performance.categorical_accuracy(y_true, y_pred)	Calculates the categorical accuracy.
performance.mcc(y_true, y_pred, **kwargs)	Calculates the Matthews correlation coefficient.
performance.f1(y_true, y_pred, **kwargs)	Calculates the F1 score.
performance.precision(y_true, y_pred, **kwargs)	Calculates the precision score.
performance.recall(y_true, y_pred, **kwargs)	Calculates the recall score.

Exceptions

exceptions.UnknownDistanceMetricError	The requested distance metric could not be located.
exceptions.UnknownPerformanceMetricError	The requested performance metric could not be located.

Random

rng.init_rng([seed])	Constructs a new random number generator.
sample.draw_random_integer(rng[, low, high])	Returns a random integer from low (inclusive) to high (exclusive).
sample.draw_random_integers(rng[, low, ...])	Returns random integers from low (inclusive) to high (exclusive).

Registry

art.load_wrapped_tensorflow_keras_classifier(...)	Loads and wraps a registered Keras classifier for compatibility with the Adversarial Robustness Toolbox.
mlflow.add_model_to_registry(active_run, ...)	Registers a trained model logged during the current run to the MLFlow registry.
mlflow.get_experiment_name(active_run)	Gets the name of the experiment for the current run.
mlflow.load_tensorflow_keras_classifier(...)	Loads a registered Keras classifier.

Tracking

mlflow.log_metrics(metrics)	Logs metrics to the MLFlow Tracking service for the current run.
mlflow.log_parameters(parameters)	Logs parameters to the MLFlow Tracking service for the current run.
mlflow.log_tensorflow_keras_estimator(...)	Logs a Keras estimator trained during the current run to the MLFlow registry.