SegmentationConcept2D

class hybrid_learning.concepts.concepts.SegmentationConcept2D(name, data, rel_size=None)

Bases: Concept

Concept with segmentation data.

Public Data Attributes:

settings	Settings dict to reproduce instance.
type	Type of the concept, which is segmentation for this sub-class.

Inherited from : py: class:Concept

type	Type of the concept, which is segmentation for this sub-class.
train_data	Training data set.
val_data	Validation data set.
test_data	Test data set.
train_val_data	Combined dataset of the defining training and validation data.
settings	Settings dict to reproduce instance.

Special Methods:

__init__(name, data[, rel_size])	Init.
__eq__(other)	In addition to __eq__() check rel_size.

Inherited from : py: class:Concept

__init__(name, data[, rel_size])	Init.
__eq__(other)	In addition to __eq__() check rel_size.
__repr__()	Nice printing function.

---

Parameters

• name (str) –

• data (DataTriple) –

• rel_size (Union[float, Tuple[float, float]]) –

__eq__(other)

In addition to __eq__() check rel_size.

Parameters

other (SegmentationConcept2D) –

__init__(name, data, rel_size=None)

Init.

For arguments name, data see Concept.

Parameters

• rel_size (Optional[Union[float, Tuple[float, float]]]) – size of the concept relative to the image size; either one value if quadratic, or a tuple of (relative width, relative height).

• name (str) –

• data (DataTriple) –

__hash__ = None

data: DataTriple

The data defining the concept, split into train/val/test data.

name: str

The name, or ID, of the concept. Should be descriptive.

rel_size: Optional[Tuple[float, float]]

Size of the concept in (width, height) relative to the image size. If set, used by detection and segmentation concept models to determine kernel size. May be None if not given (e.g. if variance too high).

property settings: Dict[str, Any]

Settings dict to reproduce instance. Use as kwargs for init.

property type: ConceptTypes

Type of the concept, which is segmentation for this sub-class.