luxonis_ml.data.augmentations.AugmentationEngine

class documentation

class AugmentationEngine(ABC):

Known subclasses: luxonis_ml.data.augmentations.AlbumentationsEngine

Constructor: AugmentationEngine(height, width, targets, n_classes, ...)

Undocumented

Method	`__init__`	Initialize augmentation pipeline from configuration.
Method	`apply`	Apply the augmentation pipeline to the data.
Property	`batch_size`	The batch size required by the augmentation pipeline.

@abstractmethod

def __init__(self, height: int, width: int, targets: Mapping[str, str], n_classes: Mapping[str, int], source_names: list[str], config: Iterable[Params], keep_aspect_ratio: bool, is_validation_pipeline: bool | None = None, pipeline_stage: Literal['train', 'val', 'test'] | None = None, min_bbox_visibility: float = 0.0, seed: int | None = None, bbox_area_threshold: float = 0.0004): ¶

overridden in luxonis_ml.data.augmentations.AlbumentationsEngine

Initialize augmentation pipeline from configuration.

Parameters
height:`int`	Target image height.
width:`int`	Target image width.
targets:`Mapping[str, str]`	Task names mapped to task types, such as `{"detection/boundingbox": "bbox"}`.
n_classes:`Mapping[str, int]`	Number of associated classes for each task, such as `{"cars/boundingbox": 2}`.
source_names:`list[str]`	Source names expected in loader images.
config:`Iterable[Params]`	Augmentation configuration. Each item describes one augmentation; interpretation is engine-specific.
keep_aspect_ratio:`bool`	Whether to preserve image aspect ratio while resizing.
is_validation_pipeline:`bool \| None`	Optional backward-compatible train-versus-eval hint. Deprecated since version 0.5.0: use `pipeline_stage` instead.
pipeline_stage:`Literal['train', 'val', 'test'] \| None`	Optional explicit pipeline stage. When provided, it takes precedence over `is_validation_pipeline`.
min_bbox_visibility:`float`	Minimum fraction of the original bounding box that must remain visible after augmentation.
seed:`int \| None`	Optional random seed for reproducible augmentation.
bbox_area_threshold:`float`	Minimum normalized area for bounding boxes to remain valid. The default removes very small boxes and their associated keypoints.

@abstractmethod

def apply(self, input_batch: list[LoaderMultiOutput]) -> LoaderMultiOutput: ¶

overridden in luxonis_ml.data.augmentations.AlbumentationsEngine

Apply the augmentation pipeline to the data.

Parameters
input_batch:`list[LoaderMultiOutput]`	Loader outputs to augment. The number of items must match the engine's batch size.

Returns
`LoaderMultiOutput`	Augmented loader output.

@property
@abstractmethod

batch_size: int = ¶

overridden in luxonis_ml.data.augmentations.AlbumentationsEngine

The batch size required by the augmentation pipeline.

The batch size is the number of images requested by the augmentation pipeline in case of batch-based augmentations.

For example, if the augmentation pipeline contains the MixUp augmentation, the batch size should be 2.

If the pipeline requires MixUp and also Mosaic4 augmentations, the batch size should be 8 = (2⋅4).