SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Add regression tests for slow sentencepiece tokenizers. (#11737)

Browse Source 
* add test_vocab_size for sentencepiece tok.

* add test_get_vocab for sentencepiece tok.

* add test_convert_token_and_id for sentencepiece tok.

* add test_tokenize_and_convert_tokens_to_string for all tok.

* improve test_tokenize_and_convert_tokens_to_string for sp. tok.

* add common tokenizer integration tests
- for albert
- for barthez

* add tokenizer integration tests to bert gen.

* add most tokenizer integration tests

* fix camembert tokenizer integration test

* add tokenizer integration test to marian

* add tokenizer integration test to reformer

* add typing and doc to tokenizer_integration_test_util

* fix tokenizer integration test of reformer

* improve test_sentencepiece_tokenize_and_convert_tokens_to_string

* empty commit to trigger CI

* fix tokenizer integration test of reformer

* remove code not needed anymore

* empty commit to trigger CI

* empty commit to trigger CI
This commit is contained in:
Philip May
2021-06-01 15:24:39 +02:00
committed by GitHub
parent c3d958b2c0
commit fcad801825
17 changed files with 624 additions and 111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

33
tests/test_tokenization_t5.py
View File

@@ -17,7 +17,7 @@ import unittest
from transformers import SPIECE_UNDERLINE, AddedToken, BatchEncoding, T5Tokenizer, T5TokenizerFast
from transformers.file_utils import cached_property, is_tf_available, is_torch_available
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers
from transformers.testing_utils import get_tests_dir, require_sentencepiece, require_tokenizers, slow
from .test_tokenization_common import TokenizerTesterMixin
@@ -48,6 +48,25 @@ class T5TokenizationTest(TokenizerTesterMixin, unittest.TestCase):
tokenizer = T5Tokenizer(SAMPLE_VOCAB)
tokenizer.save_pretrained(self.tmpdirname)
def test_convert_token_and_id(self):
"""Test ``_convert_token_to_id`` and ``_convert_id_to_token``."""
token = "<s>"
token_id = 1
self.assertEqual(self.get_tokenizer()._convert_token_to_id(token), token_id)
self.assertEqual(self.get_tokenizer()._convert_id_to_token(token_id), token)
def test_get_vocab(self):
vocab_keys = list(self.get_tokenizer().get_vocab().keys())
self.assertEqual(vocab_keys[0], "<unk>")
self.assertEqual(vocab_keys[1], "<s>")
self.assertEqual(vocab_keys[-1], "<pad>")
self.assertEqual(len(vocab_keys), 1_101)
def test_vocab_size(self):
self.assertEqual(self.get_tokenizer().vocab_size, 1_100)
def test_full_tokenizer(self):
tokenizer = T5Tokenizer(SAMPLE_VOCAB)
@@ -274,3 +293,15 @@ class T5TokenizationTest(TokenizerTesterMixin, unittest.TestCase):
self.assertTrue(special_token_id in p_output)
self.assertTrue(special_token_id in r_output)
self.assertTrue(special_token_id in cr_output)
@slow
def test_tokenizer_integration(self):
# fmt: off
expected_encoding = {'input_ids': [[31220, 7, 41, 14034, 801, 38, 3, 102, 63, 17, 127, 524, 18, 7031, 2032, 277, 11, 3, 102, 63, 17, 127, 524, 18, 2026, 17, 10761, 18, 7041, 61, 795, 879, 18, 19681, 4648, 7, 41, 12920, 382, 6, 350, 6383, 4949, 6, 2158, 12920, 382, 9, 6, 3, 4, 11160, 6, 2043, 17153, 279, 49, 17, 6, 3, 4, 434, 9688, 11439, 21, 6869, 10509, 17725, 41, 567, 9138, 61, 11, 6869, 10509, 11946, 41, 18207, 517, 61, 28, 147, 3538, 1220, 7140, 10761, 2250, 16, 910, 1220, 8024, 11, 1659, 1413, 32, 883, 2020, 344, 2215, 226, 6, 12901, 382, 127, 524, 11, 4738, 7, 127, 15390, 5, 1], [272, 24203, 19, 876, 12, 554, 18, 9719, 1659, 2647, 26352, 6497, 7, 45, 73, 9339, 400, 26, 1499, 57, 22801, 10760, 30, 321, 646, 11, 269, 2625, 16, 66, 7500, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [37, 1704, 4216, 3, 20400, 4418, 7, 147, 8, 19743, 1782, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]} # noqa: E501
# fmt: on
self.tokenizer_integration_test_util(
expected_encoding=expected_encoding,
model_name="t5-base",
revision="5a7ff2d8f5117c194c7e32ec1ccbf04642cca99b",
)