PyTorchObjectDetection

class PyTorchObjectDetection

Bases: PyTorchLearnerBackend

__init__(pipeline_cfg: RVPipelineConfig, learner_cfg: LearnerConfig, tmp_dir: str)

Parameters:

• pipeline_cfg (RVPipelineConfig) –

• learner_cfg (LearnerConfig) –

• tmp_dir (str) –

Methods

__init__(pipeline_cfg, learner_cfg, tmp_dir)
chip_dataset(dataset, chip_options[, ...])	Create and write chips for scenes in a DatasetConfig.
chip_pytorch_dataset(dataset, sample_writer, ...)
get_sample_writer()	Returns a SampleWriter for this Backend.
load_model([uri])	Load the model in preparation for one or more prediction calls.
predict_scene(scene, predict_options)	Return predictions for an entire scene using the model.
train([source_bundle_uri])	Train a model.

__init__(pipeline_cfg: RVPipelineConfig, learner_cfg: LearnerConfig, tmp_dir: str)

Parameters:

• pipeline_cfg (RVPipelineConfig) –

• learner_cfg (LearnerConfig) –

• tmp_dir (str) –

chip_dataset(dataset: DatasetConfig, chip_options: ChipOptions, dataloader_kw: dict = {}) → None

Create and write chips for scenes in a DatasetConfig.

Parameters:

• scenes – Scenes to chip.

• dataset (DatasetConfig) –

• chip_options (ChipOptions) –

• dataloader_kw (dict) –

Return type:

None

chip_pytorch_dataset(dataset: Dataset, sample_writer: PyTorchLearnerSampleWriter, chip_options: ChipOptions, split: str | None = None, dataloader_kw: dict = {}) → None

Parameters:

• dataset (Dataset) –

• sample_writer (PyTorchLearnerSampleWriter) –

• chip_options (ChipOptions) –

• split (str | None) –

• dataloader_kw (dict) –

Return type:

None

get_sample_writer()

Returns a SampleWriter for this Backend.

load_model(uri: str | None = None)

Load the model in preparation for one or more prediction calls.

Parameters:

uri (str | None) – Optional URI to load the model from.

predict_scene(scene: Scene, predict_options: ObjectDetectionPredictOptions) → ObjectDetectionLabels

Return predictions for an entire scene using the model.

Parameters:

• scene (Scene) – Scene to run inference on.

• predict_options (ObjectDetectionPredictOptions) – Prediction options.

Returns:

Labels object containing predictions

Return type:

ObjectDetectionLabels

train(source_bundle_uri=None)

Train a model.

This should download chips created by the SampleWriter, train the model, and then saving it to disk.