fairseq/fairseq/data/fairseq_dataset.py

# Copyright (c) 2017-present, Facebook, Inc.
# All rights reserved.
#
# This source code is licensed under the license found in the LICENSE file in
# the root directory of this source tree. An additional grant of patent rights
# can be found in the PATENTS file in the same directory.

import torch.utils.data


class FairseqDataset(torch.utils.data.Dataset):
    """A dataset that provides helpers for batching."""

    def __getitem__(self, index):
        raise NotImplementedError

    def __len__(self):
        raise NotImplementedError

    def collater(self, samples):
        """Merge a list of samples to form a mini-batch.

        Args:
            samples (List[int]): sample indices to collate

        Returns:
            dict: a mini-batch suitable for forwarding with a Model
        """
        raise NotImplementedError

    def get_dummy_batch(self, num_tokens, max_positions):
        """Return a dummy batch with a given number of tokens."""
        raise NotImplementedError

    def num_tokens(self, index):
        """Return the number of tokens in a sample. This value is used to
        enforce ``--max-tokens`` during batching."""
        raise NotImplementedError

    def size(self, index):
        """Return an example's size as a float or tuple. This value is used when
        filtering a dataset with ``--max-positions``."""
        raise NotImplementedError

    def ordered_indices(self):
        """Return an ordered list of indices. Batches will be constructed based
        on this order."""
        raise NotImplementedError

    @property
    def supports_prefetch(self):
        """Whether this dataset supports prefetching."""
        return False

    def prefetch(self, indices):
        """Prefetch the data required for this epoch."""
        raise NotImplementedError