Train
- module Train
- TrainWindow : public QWidget
The TrainWindow class is a PySide2 Graphical User Interface (GUI) window that is called by MainWindow class in order to configure a custom training session and initiates training for a selected Precision-Level.
Public Functions
- __init__(self, debug=False)
The constructor. Sets the size of the window and configurations for a training session. Calls setButtons function to populate window with button.
- setButtons(self)
A Mutator function that defines all buttons in TrainWindow.
- setP2(self)
A function that is triggered by the button labelled, P2.
- setP3(self)
A function that is triggered by the button labelled, P3.
- setModel(self, index)
A function that is triggered by the DropDown Menu option, Model.
- setLabelList(self)
A function that is triggered by Choose Label List button.
- setDataset(self)
A function that is triggered by Choose Dataset button.
- setMaxIteration(self)
- setCheckPointPeriod(self)
- setTestPeriod(self)
- setSteps(self)
- runLabelme(self)
A function that is triggered by Label Dataset button.
- validateTraining(self)
A Mutator function that evaulates necessary boolean flags that are all required to allow proper training given a certain requested precision level.
- validateDataset(self, new_filepath_to_dataset)
- startTraining(self)
- updateBeforeStartingTraining(self)
A function that is triggered by the button labelled, Train.
- conformDatasetToCOCO(self)
A function that is triggered by Generate Dataset button.
- populateModelSelector(self)
A Mutator function that populates the DropDown Menu labelled, Choose Model with all available pretrained models from PyTorch model zoo.
- connectTrainingButton(self)
A Mutator function that allows the Train button, to be used by the user.
- disconnectTrainingButton(self)
A Mutator function that disallows the Train button, to be used by the user.
Public Members
- train_logger
- debug
- model_name
- buttonConnected
- label_process
- label_train_process
- label_val_process
- max_iteration
- checkpoint_period
- test_period
- steps
- p2_button
- p3_button
- model_selector
- label_button
- generate_button
- validate_button
- dataset_button
- train_button
- list_button
- training_config_label
- maxiter_button
- checkpointp_button
- steps_button
- testp_button