Expose BacktranslationDataset from fairseq.data (#324)

Summary:
Pull Request resolved: https://github.com/pytorch/fairseq/pull/324

BacktranslationDataset was introduced recently but was not exposed as part of the fairseq.data module

Reviewed By: liezl200

Differential Revision: D10412717

fbshipit-source-id: 8a9d4ecd43fd376e895c450d00e765a869c95eff

fairseq/data/__init__.py

@@ -10,6 +10,7 @@ from .fairseq_dataset import FairseqDataset
 from .concat_dataset import ConcatDataset
 from .indexed_dataset import IndexedDataset, IndexedCachedDataset, IndexedInMemoryDataset, IndexedRawTextDataset
 from .append_eos_dataset import AppendEosDataset
+from .backtranslation_dataset import BacktranslationDataset
 from .language_pair_dataset import LanguagePairDataset
 from .monolingual_dataset import MonolingualDataset
 from .round_robin_zip_datasets import RoundRobinZipDatasets
@@ -39,4 +40,5 @@ __all__ = [
     'RoundRobinZipDatasets',
     'ShardedIterator',
     'TokenBlockDataset',
+    'BacktranslationDataset',
 ]