scripts package

Subpackages

Submodules

scripts.attention_unet module

Attention-based U-Net on residual frame information.

class scripts.attention_unet.ResidualTwoPlusOneDataModule(data_dir: str = 'data/train_val/', test_dir: str = 'data/test/', indices_dir: str = 'data/indices/', batch_size: int = 2, frames: int = 5, image_size: int | Tuple[int, int] = (224, 224), select_frame_method: Literal['consecutive', 'specific'] = 'specific', classification_mode: ClassificationMode = ClassificationMode.MULTICLASS_MODE, residual_mode: ResidualMode = ResidualMode.SUBTRACT_NEXT_FRAME, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.RGB, combine_train_val: bool = False, augment: bool = False, dummy_predict: DummyPredictMode = DummyPredictMode.NONE, histogram_equalize: bool = False)

Bases: LightningDataModule

Datamodule for the Residual TwoPlusOne dataset.

setup(stage)

Set up datamodule components.

train_dataloader()

An iterable or collection of iterables specifying training samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.attention_unet.ResidualAttentionCLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for Residual Attention task.

before_instantiate_classes() → None

Run some code before instantiating the classes.

Sets the torch multiprocessing mode depending on the optical flow method.

add_arguments_to_parser(parser: LightningArgumentParser)

Add extra arguments to CLI parser.

scripts.cine module

Cine Baseline model training script.

class scripts.cine.CineBaselineDataModule(frames: int = 30, data_dir: str = 'data/train_val/', test_dir: str = 'data/test/', indices_dir: str = 'data/indices/', batch_size: int = 8, image_size: int | Tuple[int, int] = (224, 224), classification_mode: ClassificationMode = ClassificationMode.MULTICLASS_MODE, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.RGB, combine_train_val: bool = False, augment: bool = False, dummy_predict: DummyPredictMode = DummyPredictMode.NONE, select_frame_method: Literal['consecutive', 'specific'] = 'consecutive')

Bases: LightningDataModule

DataModule for the Cine baseline implementation.

setup(stage)

Called at the beginning of fit (train + validate), validate, test, or predict. This is a good hook when you need to build models dynamically or adjust something about them. This hook is called on every process when using DDP.

Parameters:

stage – either 'fit', 'validate', 'test', or 'predict'

Example:

class LitModel(...):
    def __init__(self):
        self.l1 = None

    def prepare_data(self):
        download_data()
        tokenize()

        # don't do this
        self.something = else

    def setup(self, stage):
        data = load_data(...)
        self.l1 = nn.Linear(28, data.num_classes)
train_dataloader()

Get the training dataloader.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.cine.CineCLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for cine CMR task.

add_arguments_to_parser(parser: LightningArgumentParser) → None

Set the default arguments and add the arguments to the parser.

scripts.four_stream module

Four stream feature fusion attention-based U-Net on residual frame information.

class scripts.four_stream.FourStreamDataModule(data_dir: str = 'data/train_val', test_dir: str = 'data/test', indices_dir: str = 'data/indices', batch_size: int = 2, frames: int = 5, image_size: int | Tuple[int, int] = (224, 224), select_frame_method: Literal['consecutive', 'specific'] = 'specific', classification_mode: ClassificationMode = ClassificationMode.BINARY_CLASS_3_MODE, residual_mode: ResidualMode = ResidualMode.SUBTRACT_NEXT_FRAME, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.GREYSCALE, combine_train_val: bool = False, augment: bool = False, dummy_predict: DummyPredictMode = DummyPredictMode.NONE, dummy_text: bool = False)

Bases: LightningDataModule

Lightning datamodule for the four stream task.

setup(stage: str)

Called at the beginning of fit (train + validate), validate, test, or predict. This is a good hook when you need to build models dynamically or adjust something about them. This hook is called on every process when using DDP.

Parameters:

stage – either 'fit', 'validate', 'test', or 'predict'

Example:

class LitModel(...):
    def __init__(self):
        self.l1 = None

    def prepare_data(self):
        download_data()
        tokenize()

        # don't do this
        self.something = else

    def setup(self, stage):
        data = load_data(...)
        self.l1 = nn.Linear(28, data.num_classes)
train_dataloader()

An iterable or collection of iterables specifying training samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.four_stream.FourStreamAttentionCLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for 4-stream task.

before_instantiate_classes() → None

Set the last checkpoint name.

add_arguments_to_parser(parser: LightningArgumentParser)

Add extra arguments to CLI parser.

scripts.lge module

LGE Baseline model training script.

class scripts.lge.LGEBaselineDataModule(data_dir: str = 'data/train_val/', test_dir: str = 'data/test/', indices_dir: str = 'data/indices/', batch_size: int = 8, image_size: int | Tuple[int, int] = (224, 224), classification_mode: ClassificationMode = ClassificationMode.MULTICLASS_MODE, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.RGB, combine_train_val: bool = False, augment: bool = False, dummy_predict: DummyPredictMode = DummyPredictMode.NONE)

Bases: LightningDataModule

LGE MRI image data module.

setup(stage)

Called at the beginning of fit (train + validate), validate, test, or predict. This is a good hook when you need to build models dynamically or adjust something about them. This hook is called on every process when using DDP.

Parameters:

stage – either 'fit', 'validate', 'test', or 'predict'

Example:

class LitModel(...):
    def __init__(self):
        self.l1 = None

    def prepare_data(self):
        download_data()
        tokenize()

        # don't do this
        self.something = else

    def setup(self, stage):
        data = load_data(...)
        self.l1 = nn.Linear(28, data.num_classes)
train_dataloader()

An iterable or collection of iterables specifying training samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.lge.LGECLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI for LGE MRI model training.

multi_frame: bool = False
add_arguments_to_parser(parser: LightningArgumentParser) → None

Set the default arguments and add the arguments to the parser.

scripts.three_stream module

Three stream feature fusion attention-based U-Net on residual frame information.

class scripts.three_stream.ThreeStreamDataModule(data_dir: str = 'data/train_val', test_dir: str = 'data/test', indices_dir: str = 'data/indices', batch_size: int = 2, frames: int = 5, image_size: int | Tuple[int, int] = (224, 224), select_frame_method: Literal['consecutive', 'specific'] = 'specific', classification_mode: ClassificationMode = ClassificationMode.BINARY_CLASS_3_MODE, residual_mode: ResidualMode = ResidualMode.SUBTRACT_NEXT_FRAME, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.GREYSCALE, combine_train_val: bool = False, augment: bool = False, dummy_predict: DummyPredictMode = DummyPredictMode.NONE, dummy_text: bool = False)

Bases: LightningDataModule

Lightning datamodule for the three stream task.

setup(stage: str)

Called at the beginning of fit (train + validate), validate, test, or predict. This is a good hook when you need to build models dynamically or adjust something about them. This hook is called on every process when using DDP.

Parameters:

stage – either 'fit', 'validate', 'test', or 'predict'

Example:

class LitModel(...):
    def __init__(self):
        self.l1 = None

    def prepare_data(self):
        download_data()
        tokenize()

        # don't do this
        self.something = else

    def setup(self, stage):
        data = load_data(...)
        self.l1 = nn.Linear(28, data.num_classes)
train_dataloader()

An iterable or collection of iterables specifying training samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.three_stream.ThreeStreamAttentionCLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for 4-stream task.

before_instantiate_classes() → None

Set the last checkpoint name.

add_arguments_to_parser(parser: LightningArgumentParser)

Add extra arguments to CLI parser.

scripts.tscsenet module

Temporal, Spatial Squeeze, and Excitation model.

class scripts.tscsenet.TSCSECLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for cine CMR TSCSE-UNet task.

add_arguments_to_parser(parser: LightningArgumentParser)

Set the default arguments and add the arguments to the parser.

scripts.two_plus_one module

Two-plus-one architecture training script.

class scripts.two_plus_one.TwoPlusOneDataModule(data_dir: str = 'data/train_val/', test_dir: str = 'data/test/', indices_dir: str = 'data/indices/', batch_size: int = 4, frames: int = 5, image_size: int | Tuple[int, int] = (224, 224), select_frame_method: Literal['consecutive', 'specific'] = 'specific', classification_mode: ClassificationMode = ClassificationMode.MULTICLASS_MODE, num_workers: int = 8, loading_mode: LoadingMode = LoadingMode.RGB, combine_train_val: bool = False, augment: bool = False)

Bases: LightningDataModule

Datamodule for the TwoPlusOne dataset for PyTorch Lightning compatibility.

setup(stage)

Called at the beginning of fit (train + validate), validate, test, or predict. This is a good hook when you need to build models dynamically or adjust something about them. This hook is called on every process when using DDP.

Parameters:

stage – either 'fit', 'validate', 'test', or 'predict'

Example:

class LitModel(...):
    def __init__(self):
        self.l1 = None

    def prepare_data(self):
        download_data()
        tokenize()

        # don't do this
        self.something = else

    def setup(self, stage):
        data = load_data(...)
        self.l1 = nn.Linear(28, data.num_classes)
train_dataloader()

An iterable or collection of iterables specifying training samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

val_dataloader()

An iterable or collection of iterables specifying validation samples.

For more information about multiple dataloaders, see this section.

The dataloader you return will not be reloaded unless you set :paramref:`~lightning.pytorch.trainer.trainer.Trainer.reload_dataloaders_every_n_epochs` to a positive integer.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Note

If you don’t need a validation dataset and a validation_step(), you don’t need to implement this method.

test_dataloader()

An iterable or collection of iterables specifying test samples.

For more information about multiple dataloaders, see this section.

For data processing use the following pattern:

  • download in prepare_data()

  • process and split in setup()

However, the above are only necessary for distributed processing.

Warning

do not assign state in prepare_data

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware. There is no need to set it yourself.

Note

If you don’t need a test dataset and a test_step(), you don’t need to implement this method.

predict_dataloader()

An iterable or collection of iterables specifying prediction samples.

For more information about multiple dataloaders, see this section.

It’s recommended that all data downloads and preparation happen in prepare_data().

Note

Lightning tries to add the correct sampler for distributed and arbitrary hardware There is no need to set it yourself.

Returns:

A torch.utils.data.DataLoader or a sequence of them specifying prediction samples.

class scripts.two_plus_one.TwoPlusOneCLI(*args, **kwargs)

Bases: I2RInternshipCommonCLI

CLI class for cine CMR 2+1 task.

add_arguments_to_parser(parser: LightningArgumentParser)

Set the default arguments and add the arguments to the parser.

scripts.urr_attention_unet module

Attention-based U-Net on residual frame information with uncertainty.

Module contents

Training/Validation/Test scripts for various tasks.