ObjectDetection3DTracer Class

ml_debugger.training.object_detection_3d.object_detection_3d_torchtracer.ObjectDetection3DTrainingTorchTracer

Bases: TrainingTracer, ObjectDetection3DTorchTracer

Base TrainingTracer for 3D object detection tasks using PyTorch models.

Subclasses must implement the following abstract methods

• detect(model): Static method to check if this subclass handles the model.
• _get_n_class(model): Extract number of classes from the model.
• _register_model_hooks(model): Register forward hooks for logit capture.
• _parse_model_output(model_output): Convert model output to a standardized Detection3DOutput.

Optionally override

• _apply_nms_config(model, nms_config): Apply NMS parameters to the model.

__init__(model, model_name, version_name, target_layers=None, additional_fields=None, additional_label_fields=None, auto_sync=False, force_table_recreate=False, api_endpoint=None, api_key=None, n_class=None, score_thresh=None, iou_thresh=None, max_detections_per_frame=None, pc_range=None, ground_truth_provider=None)

Initialize 3D object detection training tracer.

Parameters:

Name	Type	Description	Default
model	Module	PyTorch model to trace.	required
model_name	str	Name of the ML model.	required
version_name	str	Version identifier for the ML model.	required
target_layers	Optional[Dict[str, str]]	Mapping of layer aliases to module paths.	None
additional_fields	Optional[List[dict]]	Extra fields for database schema.	None
additional_label_fields	Optional[List[dict]]	Extra fields for groundtruth database schema.	None
auto_sync	bool	Enable background syncing of logged data.	False
force_table_recreate	bool	Whether to drop and recreate existing tables.	False
api_endpoint	Optional[str]	URL of the service API for data upload.	None
api_key	Optional[str]	API key for authenticating with the service.	None
n_class	Optional[int]	Number of classes. Auto-detected from model if None.	None
score_thresh	Optional[float]	Minimum score threshold for detections. None to infer from the model's test_cfg.	None
iou_thresh	Optional[float]	BEV IoU threshold for NMS. None to infer from the model's test_cfg.	None
max_detections_per_frame	Optional[int]	Maximum detections per frame after NMS. None to infer from the model's test_cfg.	None
pc_range	Optional[List[float]]	Point cloud range [x_min, y_min, z_min, x_max, y_max, z_max].	None
ground_truth_provider	Optional[Any]	Optional provider implementing resolve(sample_tokens: List[str]) -> List[Dict[str, Any]]. Used when ground_truth is omitted at call time.	None

__call__(model_input, ground_truth, input_ids, dataset_type='train', n_epoch=None, **kwargs)

Invoke the tracer on a single inference, recording I/O data.

Parameters:

Name	Type	Description	Default
model_input	Any	Input data for the model. Can be a torch.Tensor (point cloud) or a dict (MMDetection3D format with points, img, img_metas).	required
ground_truth	Optional[Any]	Ground truth 3D bounding boxes and labels. If omitted, ground_truth_provider must be configured. Accepts List[Dict] or Dict[str, List] format with keys: 'boxes_3d' (N,7), 'labels_3d' (N,), optional 'velocities' (N,2), or MMDet3D-style dict with gt_bboxes_3d / gt_labels_3d.	required
input_ids	List[str]	Identifiers of each input data.	required
dataset_type	str	Identifier of input dataset.	'train'
n_epoch	Optional[int]	Number identifier of training iteration.	None
**kwargs	Any	Additional keyword arguments passed to _parse_and_save_io_data.	{}

Returns:

Type	Description
Any	The raw output produced by the model.

get_hooked_features(layer_name)

Retrieve the captured output for a given layer alias.

Parameters:

Name	Type	Description	Default
layer_name	str	Alias of the layer whose activation was captured.	required

Returns:

Name	Type	Description
Any	Any	Activation data stored for the specified layer.

Raises:

Type	Description
KeyError	If no activation has been captured for layer_name.

export(output_path=None, n_epoch=None)

Export extracted features and ground truth data into a ZIP archive.

Parameters:

Name	Type	Description	Default
output_path	Optional[str]	Path or directory for saving the ZIP file. If no .zip extension, the default filename is appended. Defaults to cwd.	None
n_epoch	Optional[int]	Target epoch number. If None, exports all epochs.	None

Returns:

Type	Description
Optional[Path]	Path to the created ZIP file, or None on non-primary distributed ranks.

wait_for_save(n_epoch=None, interval=3)

upload(n_epoch=None)