lambda: algorithmic_math.algebra_inverse(26, 3, 3, 10000),
lambda: None), # test set
"parsing_english_ptb8k":
(lambda: wsj_parsing.parsing_token_generator(FLAGS.data_dir, FLAGS.tmp_dir,
True, 2**13, 2**9),
lambda: wsj_parsing.parsing_token_generator(FLAGS.data_dir, FLAGS.tmp_dir,
False, 2**13, 2**9),
lambda: None), # test set
"parsing_english_ptb16k":
(lambda: wsj_parsing.parsing_token_generator(FLAGS.data_dir, FLAGS.tmp_dir,
True, 2**14, 2**9),
lambda: wsj_parsing.parsing_token_generator(FLAGS.data_dir, FLAGS.tmp_dir,
False, 2**14, 2**9),
lambda: None), # test set
"inference_snli32k":
(lambda: snli.snli_token_generator(FLAGS.tmp_dir, True, 2**15),
lambda: snli.snli_token_generator(FLAGS.tmp_dir, False, 2**15),
lambda: None), # test set
"audio_timit_characters_test":
(lambda: audio.timit_generator(FLAGS.data_dir, FLAGS.tmp_dir, True, 1718),
lambda: audio.timit_generator(FLAGS.data_dir, FLAGS.tmp_dir, False, 626),
lambda: None), # test set
"audio_timit_tokens_8k_test":
(lambda: audio.timit_generator(FLAGS.data_dir,
FLAGS.tmp_dir,
True,
1718,
vocab_filename="vocab.endefr.%d" % 2**13,
vocab_size=2**13),
lambda: audio.timit_generator(FLAGS.data_dir,
FLAGS.tmp_dir,
_SUPPORTED_PROBLEM_GENERATORS = {
"algorithmic_algebra_inverse": (
lambda: algorithmic_math.algebra_inverse(26, 0, 2, 100000),
lambda: algorithmic_math.algebra_inverse(26, 3, 3, 10000)),
"parsing_english_ptb8k": (
lambda: wsj_parsing.parsing_token_generator(
FLAGS.data_dir, FLAGS.tmp_dir, True, 2**13, 2**9),
lambda: wsj_parsing.parsing_token_generator(
FLAGS.data_dir, FLAGS.tmp_dir, False, 2**13, 2**9)),
"parsing_english_ptb16k": (
lambda: wsj_parsing.parsing_token_generator(
FLAGS.data_dir, FLAGS.tmp_dir, True, 2**14, 2**9),
lambda: wsj_parsing.parsing_token_generator(
FLAGS.data_dir, FLAGS.tmp_dir, False, 2**14, 2**9)),
"inference_snli32k": (
lambda: snli.snli_token_generator(FLAGS.tmp_dir, True, 2**15),
lambda: snli.snli_token_generator(FLAGS.tmp_dir, False, 2**15),
),
"audio_timit_characters_test": (
lambda: audio.timit_generator(
FLAGS.data_dir, FLAGS.tmp_dir, True, 1718),
lambda: audio.timit_generator(
FLAGS.data_dir, FLAGS.tmp_dir, False, 626)),
"audio_timit_tokens_8k_test": (
lambda: audio.timit_generator(
FLAGS.data_dir, FLAGS.tmp_dir, True, 1718,
vocab_filename="vocab.endefr.%d" % 2**13, vocab_size=2**13),
lambda: audio.timit_generator(
FLAGS.data_dir, FLAGS.tmp_dir, False, 626,
vocab_filename="vocab.endefr.%d" % 2**13, vocab_size=2**13)),
"audio_timit_tokens_32k_test": (
