Managing Data

Data Containers in Lightning

There are a few different data containers used in Lightning:

Data objects




The PyTorch Dataset represents a map from keys to data samples.


The PyTorch IterableDataset represents a stream of data.


The PyTorch DataLoader represents a Python iterable over a Dataset.


A LightningDataModule is simply a collection of: training DataLoader(s), validation DataLoader(s), test DataLoader(s) and predict DataLoader(s), along with the matching transforms and data processing/downloads steps required.

Why Use LightningDataModule?

The LightningDataModule was designed as a way of decoupling data-related hooks from the LightningModule so you can develop dataset agnostic models. The LightningDataModule makes it easy to hot swap different Datasets with your model, so you can test it and benchmark it across domains. It also makes sharing and reusing the exact data splits and transforms across projects possible.

Read this for more details on LightningDataModule.

Multiple Datasets

There are a few ways to pass multiple Datasets to Lightning:

  1. Create a DataLoader that iterates over multiple Datasets under the hood.

  2. In the training loop, you can pass multiple DataLoaders as a dict or list/tuple, and Lightning will automatically combine the batches from different DataLoaders.

  3. In the validation, test, or prediction, you have the option to return multiple DataLoaders as list/tuple, which Lightning will call sequentially or combine the DataLoaders using CombinedLoader, which Lightning will automatically combine the batches from different DataLoaders.

Using LightningDataModule

You can set more than one DataLoader in your LightningDataModule using its DataLoader hooks and Lightning will use the correct one.

class DataModule(LightningDataModule):

    def train_dataloader(self):
        return DataLoader(self.train_dataset)

    def val_dataloader(self):
        return [DataLoader(self.val_dataset_1), DataLoader(self.val_dataset_2)]

    def test_dataloader(self):
        return DataLoader(self.test_dataset)

    def predict_dataloader(self):
        return DataLoader(self.predict_dataset)

Using LightningModule Hooks

Concatenated Dataset

For training with multiple Datasets, you can create a DataLoader class which wraps your multiple Datasets using ConcatDataset. This, of course, also works for testing, validation, and prediction Datasets.

from import ConcatDataset

class LitModel(LightningModule):
    def train_dataloader(self):
        concat_dataset = ConcatDataset(datasets.ImageFolder(traindir_A), datasets.ImageFolder(traindir_B))

        loader = DataLoader(
            concat_dataset, batch_size=args.batch_size, shuffle=True, num_workers=args.workers, pin_memory=True
        return loader

    def val_dataloader(self):
        # SAME

    def test_dataloader(self):
        # SAME

Return Multiple DataLoaders

You can set multiple DataLoaders in your LightningModule, and Lightning will take care of batch combination.

For more details, refer to multiple_trainloader_mode

class LitModel(LightningModule):
    def train_dataloader(self):
        loader_a = DataLoader(range(6), batch_size=4)
        loader_b = DataLoader(range(15), batch_size=5)

        # pass loaders as a dict. This will create batches like this:
        # {'a': batch from loader_a, 'b': batch from loader_b}
        loaders = {"a": loader_a, "b": loader_b}

        # OR:
        # pass loaders as sequence. This will create batches like this:
        # [batch from loader_a, batch from loader_b]
        loaders = [loader_a, loader_b]

        return loaders

Furthermore, Lightning also supports nested lists and dicts (or a combination).

class LitModel(LightningModule):
    def train_dataloader(self):
        loader_a = DataLoader(range(8), batch_size=4)
        loader_b = DataLoader(range(16), batch_size=2)

        return {"a": loader_a, "b": loader_b}

    def training_step(self, batch, batch_idx):
        # access a dictionary with a batch from each DataLoader
        batch_a = batch["a"]
        batch_b = batch["b"]
class LitModel(LightningModule):
    def train_dataloader(self):
        loader_a = DataLoader(range(8), batch_size=4)
        loader_b = DataLoader(range(16), batch_size=4)
        loader_c = DataLoader(range(32), batch_size=4)
        loader_c = DataLoader(range(64), batch_size=4)

        # pass loaders as a nested dict. This will create batches like this:
        loaders = {"loaders_a_b": [loader_a, loader_b], "loaders_c_d": {"c": loader_c, "d": loader_d}}
        return loaders

    def training_step(self, batch, batch_idx):
        # access the data
        batch_a_b = batch["loaders_a_b"]
        batch_c_d = batch["loaders_c_d"]

        batch_a = batch_a_b[0]
        batch_b = batch_a_b[1]

        batch_c = batch_c_d["c"]
        batch_d = batch_c_d["d"]

Alternatively, you can also pass in a CombinedLoader containing multiple DataLoaders.

from pytorch_lightning.trainer.supporters import CombinedLoader

def train_dataloader(self):
    loader_a = DataLoader()
    loader_b = DataLoader()
    loaders = {"a": loader_a, "b": loader_b}
    combined_loader = CombinedLoader(loaders, mode="max_size_cycle")
    return combined_loader

def training_step(self, batch, batch_idx):
    batch_a = batch["a"]
    batch_b = batch["b"]

Multiple Validation/Test/Predict DataLoaders

For validation, test and predict DataLoaders, you can pass a single DataLoader or a list of them. This optional named parameter can be used in conjunction with any of the above use cases. You can choose to pass the batches sequentially or simultaneously, as is done for the training step. The default mode for these DataLoaders is sequential. Note that when using a sequence of DataLoaders you need to add an additional argument dataloader_idx in their corresponding step specific hook. The corresponding loop will process the DataLoaders in sequential order; that is, the first DataLoader will be processed completely, then the second one, and so on.

Refer to the following for more details for the default sequential option:

def val_dataloader(self):
    loader_1 = DataLoader()
    loader_2 = DataLoader()
    return [loader_1, loader_2]

def validation_step(self, batch, batch_idx, dataloader_idx):

Evaluation DataLoaders are iterated over sequentially. If you want to iterate over them in parallel, PyTorch Lightning provides a CombinedLoader object which supports collections of DataLoaders such as list, tuple, or dictionary. The DataLoaders can be accessed using in the same way as the provided structure:

from pytorch_lightning.trainer.supporters import CombinedLoader

def val_dataloader(self):
    loader_a = DataLoader()
    loader_b = DataLoader()
    loaders = {"a": loader_a, "b": loader_b}
    combined_loaders = CombinedLoader(loaders, mode="max_size_cycle")
    return combined_loaders

def validation_step(self, batch, batch_idx):
    batch_a = batch["a"]
    batch_b = batch["b"]

Evaluate with Additional DataLoaders

You can evaluate your models using additional DataLoaders even if the DataLoader specific hooks haven’t been defined within your LightningModule. For example, this would be the case if your test data set is not available at the time your model was declared. Simply pass the test set to the test() method:

# setup your DataLoader
test = DataLoader(...)

# test (pass in the loader)

Accessing DataLoaders within LightningModule

In the case that you require access to the DataLoader or Dataset objects, DataLoaders for each step can be accessed using the Trainer object:

from pytorch_lightning import LightningModule

class Model(LightningModule):
    def test_step(self, batch, batch_idx, dataloader_idx):
        test_dl = self.trainer.test_dataloaders[dataloader_idx]
        test_dataset = test_dl.dataset
        test_sampler = test_dl.sampler
        # extract metadata, etc. from the dataset:

If you are using a CombinedLoader object which allows you to fetch batches from a collection of DataLoaders simultaneously which supports collections of DataLoader such as list, tuple, or dictionary. The DataLoaders can be accessed using the same collection structure:

from pytorch_lightning.trainer.supporters import CombinedLoader

test_dl1 = ...
test_dl2 = ...

# If you provided a list of DataLoaders:

combined_loader = CombinedLoader([test_dl1, test_dl2])
list_of_loaders = combined_loader.loaders
test_dl1 = list_of_loaders.loaders[0]

# If you provided dictionary of DataLoaders:

combined_loader = CombinedLoader({"dl1": test_dl1, "dl2": test_dl2})
dictionary_of_loaders = combined_loader.loaders
test_dl1 = dictionary_of_loaders["dl1"]

Sequential Data

Lightning has built in support for dealing with sequential data.

Packed Sequences as Inputs

When using PackedSequence, do two things:

  1. Return either a padded tensor in dataset or a list of variable length tensors in the DataLoader’s collate_fn (example shows the list implementation).

  2. Pack the sequence in forward or training and validation steps depending on use case.

# For use in DataLoader
def collate_fn(batch):
    x = [item[0] for item in batch]
    y = [item[1] for item in batch]
    return x, y

# In LightningModule
def training_step(self, batch, batch_idx):
    x = rnn.pack_sequence(batch[0], enforce_sorted=False)
    y = rnn.pack_sequence(batch[1], enforce_sorted=False)

Truncated Backpropagation Through Time (TBPTT)

There are times when multiple backwards passes are needed for each batch. For example, it may save memory to use Truncated Backpropagation Through Time when training RNNs.

Lightning can handle TBPTT automatically via this flag.

from pytorch_lightning import LightningModule

class MyModel(LightningModule):
    def __init__(self):
        # Important: This property activates truncated backpropagation through time
        # Setting this value to 2 splits the batch into sequences of size 2
        self.truncated_bptt_steps = 2

    # Truncated back-propagation through time
    def training_step(self, batch, batch_idx, hiddens):
        # the training step must be updated to accept a ``hiddens`` argument
        # hiddens are the hiddens from the previous truncated backprop step
        out, hiddens = self.lstm(data, hiddens)
        return {"loss": ..., "hiddens": hiddens}


If you need to modify how the batch is split, override tbptt_split_batch().

Iterable Datasets

Lightning supports using IterableDataset as well as map-style Datasets. IterableDatasets provide a more natural option when using sequential data.


When using an IterableDataset you must set the val_check_interval to 1.0 (the default) or an int (specifying the number of training batches to run before each validation loop) when initializing the Trainer. This is because the IterableDataset does not have a __len__ and Lightning requires this to calculate the validation interval when val_check_interval is less than one. Similarly, you can set limit_{mode}_batches to a float or an int. If it is set to 0.0 or 0, it will set num_{mode}_batches to 0, if it is an int, it will set num_{mode}_batches to limit_{mode}_batches, if it is set to 1.0 it will run for the whole dataset, otherwise it will throw an exception. Here mode can be train/val/test/predict.

When iterable datasets are used, Lightning will pre-fetch 1 batch (in addition to the current batch) so it can detect when the training will stop and run validation if necessary.

# IterableDataset
class CustomDataset(IterableDataset):
    def __init__(self, data):
        self.data_source = data

    def __iter__(self):
        return iter(self.data_source)

# Setup DataLoader
def train_dataloader(self):
    seq_data = ["A", "long", "time", "ago", "in", "a", "galaxy", "far", "far", "away"]
    iterable_dataset = CustomDataset(seq_data)

    dataloader = DataLoader(dataset=iterable_dataset, batch_size=5)
    return dataloader
# Set val_check_interval
trainer = Trainer(val_check_interval=100)

# Set limit_val_batches to 0.0 or 0
trainer = Trainer(limit_val_batches=0.0)

# Set limit_val_batches as an int
trainer = Trainer(limit_val_batches=100)

© Copyright Copyright (c) 2018-2023, Lightning AI et al...

Built with Sphinx using a theme provided by Read the Docs.