def prepare_option_infos(self) -> List[OptionInfo]:
model = OptionInfo(name='Model',
direction=OptionDirection.OUTPUT,
values=['MODEL'])
classes = OptionInfo(name='Class',
direction=OptionDirection.PARAMETER,
values=['10'])
return [model, classes]
def prepare_option_infos(cls):
model = OptionInfo(
name='Model',
direction=OptionDirection.OUTPUT,
values=['MODEL'])
size = OptionInfo(
name='Size',
direction=OptionDirection.PARAMETER,
values=['28'])
channel = OptionInfo(
name='Channel',
direction=OptionDirection.PARAMETER,
values=['1'])
return [model, size, channel]
def prepare_option_infos(self) -> List[OptionInfo]:
model = OptionInfo(name='Model',
direction=OptionDirection.OUTPUT,
values=['MODEL'])
size = OptionInfo(name='Size',
direction=OptionDirection.PARAMETER,
values=['28'])
channel = OptionInfo(name='Channel',
direction=OptionDirection.PARAMETER,
values=['1'])
cls = OptionInfo(name='Class',
direction=OptionDirection.PARAMETER,
values=['10'])
return [model, size, channel, cls]
def prepare_option_infos(self) -> List[OptionInfo]:
src = OptionInfo(name='SrcModel',
direction=OptionDirection.INPUT,
values=['MODEL'])
dest = OptionInfo(name='DestModel',
direction=OptionDirection.OUTPUT,
values=['MODEL'])
optimizer = OptionInfo(name='Optimizer',
direction=OptionDirection.PARAMETER,
values=['SGD,Adam'])
lr = OptionInfo(name='LearningRate',
direction=OptionDirection.PARAMETER,
values=['0.01'])
return [src, dest, optimizer, lr]
def prepare_option_infos(self) -> List[OptionInfo]:
src = OptionInfo(name='ModelFile',
direction=OptionDirection.INPUT,
values=['MODEL_FILE'])
data = OptionInfo(name='TestDataset',
direction=OptionDirection.INPUT,
values=['TEST_DATASET'])
size = OptionInfo(name='TestDatasetSize',
direction=OptionDirection.INPUT,
values=['TEST_DATASET_SIZE'])
prediction_type = OptionInfo(name='Prediction',
direction=OptionDirection.PARAMETER,
values=['Multiple', 'Binarization'])
bin_class = OptionInfo(name='Prediction.BinarizationClass',
direction=OptionDirection.PARAMETER,
values=['0'])
bin_threshold = OptionInfo(name='Prediction.BinarizationThreshold',
direction=OptionDirection.PARAMETER,
values=['0.5'])
return [src, data, size, prediction_type, bin_class, bin_threshold]
def prepare_option_infos(self) -> List[OptionInfo]:
model = OptionInfo(name='Model',
direction=OptionDirection.INPUT,
values=['MODEL'])
train = OptionInfo(name='TrainDataset',
direction=OptionDirection.INPUT,
values=['TRAIN_DATASET'])
train_size = OptionInfo(name='TrainDatasetSize',
direction=OptionDirection.INPUT,
values=['TRAIN_DATASET_SIZE'])
validation = OptionInfo(name='ValidaitonDataset',
direction=OptionDirection.INPUT,
values=['VALIDATION_DATASET'])
validation_size = OptionInfo(name='ValidationDatasetSize',
direction=OptionDirection.INPUT,
values=['VALIDATION_DATASET_SIZE'])
tensorboard = OptionInfo(name='TensorBoard',
direction=OptionDirection.PARAMETER,
values=['True', 'False'])
trial_name = OptionInfo(name='TraialName',
direction=OptionDirection.PARAMETER,
values=['traial1'])
early_stopping = OptionInfo(name='EarlyStopping',
direction=OptionDirection.PARAMETER,
values=['True', 'False'])
early_stopping_patience = OptionInfo(
name='EarlyStopping.Patience',
direction=OptionDirection.PARAMETER,
values=['5'])
early_stopping_monitor = OptionInfo(
name='EarlyStopping.Monitor',
direction=OptionDirection.PARAMETER,
values=['loss', 'val_loss'])
check_point = OptionInfo(name='ModelCheckpoint',
direction=OptionDirection.PARAMETER,
values=['True', 'False'])
check_point_save_best = OptionInfo(name='ModelCheckpoint.SaveBestOnly',
direction=OptionDirection.PARAMETER,
values=['True', 'False'])
check_point_save_weigths_only = OptionInfo(
name='ModelCheckpoint.SaveWeightsOnly',
direction=OptionDirection.PARAMETER,
values=['True', 'False'])
check_point_file_path = OptionInfo(name='ModelCheckpoint.FilePath',
direction=OptionDirection.PARAMETER,
values=['weights.hdf5'])
# values=['weights-{epoch:02d}-{val_loss:.2f}.hdf5'])
max_epoch = OptionInfo(name='MaxEpoch',
direction=OptionDirection.PARAMETER,
values=['10'])
batch = OptionInfo(name='BatchSize',
direction=OptionDirection.PARAMETER,
values=['100'])
return [
model, train, train_size, validation, validation_size, tensorboard,
trial_name, early_stopping, early_stopping_patience,
early_stopping_monitor, check_point, check_point_save_best,
check_point_save_weigths_only, check_point_file_path, max_epoch,
batch
]