Tasks#
Task Initialization of OTX Action Task.
- class otx.algorithms.action.tasks.ActionInferenceTask(task_environment: TaskEnvironment, **kwargs)#
Inference Task Implementation of OTX Action Task.
- evaluate(output_resultset: ResultSetEntity)#
Evaluate function of OTX Action Task.
- export(export_type: ExportType, output_model: ModelEntity, precision: ModelPrecision = ModelPrecision.FP32, dump_features: bool = False)#
Export function of OTX Action Task.
- infer(dataset: DatasetEntity, inference_parameters: Optional[InferenceParameters] = None) DatasetEntity #
Main infer function of OTX Action Task.
- unload()#
Unload the task.
- class otx.algorithms.action.tasks.ActionOpenVINOTask(task_environment: TaskEnvironment)#
Task implementation for OTX Action Recognition using OpenVINO backend.
- deploy(output_model: ModelEntity) None #
Deploy function of OpenVINOTask.
- evaluate(output_resultset: ResultSetEntity, evaluation_metric: Optional[str] = None)#
Evaluate function of OpenVINOTask.
- infer(dataset: DatasetEntity, inference_parameters: Optional[InferenceParameters] = None) DatasetEntity #
Infer function of OpenVINOTask for Action Recognition.
- load_inferencer() ActionOpenVINOInferencer #
load_inferencer function of OpenVINOTask for Action Recognition.
- optimize(optimization_type: OptimizationType, dataset: DatasetEntity, output_model: ModelEntity, optimization_parameters: Optional[OptimizationParameters] = None)#
Optimize function of OpenVINOTask.
- class otx.algorithms.action.tasks.ActionTrainTask(task_environment: TaskEnvironment, **kwargs)#
Train Task Implementation of OTX Action Task.
- cancel_training()#
Cancel training function in ActionTrainTask.
Sends a cancel training signal to gracefully stop the optimizer. The signal consists of creating a ‘.stop_training’ file in the current work_dir. The runner checks for this file periodically. The stopping mechanism allows stopping after each iteration, but validation will still be carried out. Stopping will therefore take some time.
- save_model(output_model: ModelEntity)#
Save best model weights in ActionTrainTask.
- train(dataset: DatasetEntity, output_model: ModelEntity, train_parameters: Optional[TrainParameters] = None)#
Train function in ActionTrainTask.