vis4d.eval.bdd100k

BDD100K evaluators.

class BDD100KDetectEvaluator(annotation_path, config_path, mask_threshold=0.0)[source]

BDD100K 2D detection evaluation class.

Initialize the evaluator.

__repr__()[source]

Concise representation of the dataset evaluator.

Return type:

str

class BDD100KSegEvaluator(annotation_path)[source]

BDD100K segmentation evaluation class.

Initialize the evaluator.

__repr__()[source]

Concise representation of the dataset evaluator.

Return type:

str

evaluate(metric)[source]

Evaluate the dataset.

Return type:

tuple[Dict[str, Union[float, int, Tensor]], str]

gather(gather_func)[source]

Gather variables in case of distributed setting (if needed).

Parameters:

gather_func (Callable[[Any], Any]) – Gather function.

Return type:

None

process_batch(data_names, masks_list)[source]

Process tracking results.

Return type:

None

reset()[source]

Reset the evaluator.

Return type:

None

property metrics: list[str]

Supported metrics.

class BDD100KTrackEvaluator(annotation_path, config_path='box_track', mask_threshold=0.0)[source]

BDD100K 2D tracking evaluation class.

Initialize the evaluator.

__repr__()[source]

Concise representation of the dataset evaluator.

Return type:

str

evaluate(metric)[source]

Evaluate the dataset.

Return type:

tuple[Dict[str, Union[float, int, Tensor]], str]

property metrics: list[str]

Supported metrics.

Modules

vis4d.eval.bdd100k.detect

BDD100K detection evaluator.

vis4d.eval.bdd100k.seg

BDD100K segmentation evaluator.

vis4d.eval.bdd100k.track

BDD100K tracking evaluator.