ObjectDetectionLogger Class

ml_debugger.monitoring.object_detection.object_detection_torchlogger.ObjectDetectionTorchLogger

Bases: MonitoringLogger, ObjectDetectionTorchTracer

Base MonitoringLogger for 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, img_sizes): Convert model output to a standardized DetectionOutput.

Optionally override

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

__init__(model, model_name, version_name, result_name=None, n_epoch='latest', target_layers=None, additional_fields=None, auto_sync=True, force_table_recreate=False, api_endpoint=None, api_key=None, n_class=None, score_thresh=None, iou_thresh=None, max_detections_per_image=None)

Initialize object detection monitoring logger.

引数：

名前	タイプ	デスクリプション	デフォルト
model	Module	PyTorch model to trace.	必須
model_name	str	Name of the ML model.	必須
version_name	str	Version identifier for the ML model.	必須
result_name	Optional[str]	The name of the existing evaluation result to retrieve.	None
n_epoch	Union[str, Optional[int]]	Filter option for n_epoch value.	'latest'
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
auto_sync	bool	Enable background syncing of logged data.	True
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 user-visible NMS decisions.	None
iou_thresh	Optional[float]	IoU threshold for user-visible NMS decisions.	None
max_detections_per_image	Optional[int]	Maximum detections per image for user-visible output.	None

__call__(model_input, input_ids=None, dataset_type='deploying', **kwargs)

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

引数：

名前	タイプ	デスクリプション	デフォルト
model_input	Any	Input data for the model inference.	必須
input_ids	Optional[List[str]]	Identifiers of each input data.	None
dataset_type	str	Identifier of input dataset. (e.g. 'train', 'val', 'test', 'deploying')	'deploying'
**kwargs	Any	Additional keyword arguments for parsing and saving I/O data.	{}

戻り値：

タイプ	デスクリプション
Any	The raw output produced by the model.

get_hooked_features(layer_name)

Retrieve the captured output for a given layer alias.

引数：

名前	タイプ	デスクリプション	デフォルト
layer_name	str	Alias of the layer whose activation was captured.	必須

戻り値：

名前	タイプ	デスクリプション
Any	Any	Activation data stored for the specified layer.

発生：

タイプ	デスクリプション
KeyError	If no activation has been captured for layer_name.

export(output_path=None)

Export extracted features into a ZIP archive.

Uses the internal n_epoch resolved during validator setup to filter records, consistent with upload() and wait_for_save().

引数：

名前	タイプ	デスクリプション	デフォルト
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

戻り値：

タイプ	デスクリプション
Optional[Path]	Path to the created ZIP file, or None on non-primary distributed ranks.

wait_for_save(interval=3)

upload()