def test_source_data_files(self):
wmt_8k = translate_ende.TranslateEndeWmt8k()
wmt_32k = translate_ende.TranslateEndeWmt32k()
eval_split = problem.DatasetSplit.EVAL
train_split = problem.DatasetSplit.TRAIN
wmt_8k_eval_files = wmt_8k.source_data_files(eval_split)
wmt_32k_eval_files = wmt_32k.source_data_files(eval_split)
self.assertListEqual(wmt_8k_eval_files, wmt_32k_eval_files)
self.assertGreater(len(wmt_8k_eval_files), 0)
wmt_8k_train_files = wmt_8k.source_data_files(train_split)
wmt_32k_train_files = wmt_32k.source_data_files(train_split)
self.assertListEqual(wmt_8k_train_files, wmt_32k_train_files)
self.assertGreater(len(wmt_8k_train_files), 0)
def test_vocab_size(self):
wmt_8k = translate_ende.TranslateEndeWmt8k()
wmt_32k = translate_ende.TranslateEndeWmt32k()
self.assertEqual(wmt_8k.approx_vocab_size, 8192)
self.assertEqual(wmt_32k.approx_vocab_size, 32768)
def test_additional_datasets(self):
wmt_8k = translate_ende.TranslateEndeWmt8k()
wmt_32k = translate_ende.TranslateEndeWmt32k()
self.assertListEqual(wmt_8k.additional_training_datasets, [])
self.assertListEqual(wmt_32k.additional_training_datasets, [])
def vocab_problem(self):
return translate_ende.TranslateEndeWmt32k()