frameworks.ultralytics.services.data.utils

utils

Functions:

Name	Description
detect_inference_type_from_experiment	Detects the inference type from the types of all attached dataset versions.
generate_data_yaml
prepare_classification_data	Prepares and organizes a dataset collection for Ultralytics classification tasks.

detect_inference_type_from_experiment(experiment)

Detects the inference type from the types of all attached dataset versions. Ensures all attached datasets are of the same type, otherwise raises an error.

Parameters:

Name	Type	Description	Default
experiment	Experiment	The experiment to inspect.	required

Returns:

Name	Type	Description
InferenceType	InferenceType	The common inference type of all attached datasets.

generate_data_yaml(dataset_collection)

prepare_classification_data(dataset_collection)

Prepares and organizes a dataset collection for Ultralytics classification tasks.

This function iterates over each dataset in the provided DatasetCollection, organizing them using the ClassificationDatasetPreparator to structure the dataset for use with Ultralytics classification. Each dataset is moved into a new directory, with the structure suitable for Ultralytics training.

Parameters:

Name	Type	Description	Default
dataset_collection	DatasetCollection	The original dataset collection to be prepared for classification.	required

Returns:

Name	Type	Description
DatasetCollection	DatasetCollection[CocoDataset]	A dataset collection where each dataset has been organized and prepared for Ultralytics classification tasks.