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

Use lru_cache for tokenization tests (#36818)

Browse Source 
* fix

* fix

* fix

* fix

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
This commit is contained in:
Yih-Dar
2025-03-28 15:09:35 +01:00
committed by GitHub
parent 3af425d4c6
commit 1fcaad6df9
92 changed files with 1301 additions and 884 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
tests/models/bert/test_tokenization_bert.py
View File

@@ -41,8 +41,9 @@ class BertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
space_between_special_tokens = True
from_pretrained_filter = filter_non_english
def setUp(self):
super().setUp()
@classmethod
def setUpClass(cls):
super().setUpClass()
vocab_tokens = [
"[UNK]",
@@ -61,8 +62,8 @@ class BertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
"low",
"lowest",
]
self.vocab_file = os.path.join(self.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
with open(self.vocab_file, "w", encoding="utf-8") as vocab_writer:
cls.vocab_file = os.path.join(cls.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
with open(cls.vocab_file, "w", encoding="utf-8") as vocab_writer:
vocab_writer.write("".join([x + "\n" for x in vocab_tokens]))
def get_input_output_texts(self, tokenizer):
@@ -257,7 +258,7 @@ class BertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
def test_offsets_with_special_characters(self):
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f"{tokenizer.__class__.__name__} ({pretrained_name})"):
tokenizer_r = self.rust_tokenizer_class.from_pretrained(pretrained_name, **kwargs)
tokenizer_r = self.get_rust_tokenizer(pretrained_name, **kwargs)
sentence = f"A, naïve {tokenizer_r.mask_token} AllenNLP sentence."
tokens = tokenizer_r.encode_plus(
@@ -312,8 +313,8 @@ class BertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
for tokenizer, pretrained_name, kwargs in self.tokenizers_list:
with self.subTest(f"{tokenizer.__class__.__name__} ({pretrained_name})"):
kwargs["tokenize_chinese_chars"] = True
tokenizer_p = self.tokenizer_class.from_pretrained(pretrained_name, **kwargs)
tokenizer_r = self.rust_tokenizer_class.from_pretrained(pretrained_name, **kwargs)
tokenizer_p = self.get_tokenizer(pretrained_name, **kwargs)
tokenizer_r = self.get_rust_tokenizer(pretrained_name, **kwargs)
ids_without_spe_char_p = tokenizer_p.encode(text_with_chinese_char, add_special_tokens=False)
ids_without_spe_char_r = tokenizer_r.encode(text_with_chinese_char, add_special_tokens=False)
@@ -326,8 +327,8 @@ class BertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
self.assertListEqual(tokens_without_spe_char_r, list_of_commun_chinese_char)
kwargs["tokenize_chinese_chars"] = False
tokenizer_r = self.rust_tokenizer_class.from_pretrained(pretrained_name, **kwargs)
tokenizer_p = self.tokenizer_class.from_pretrained(pretrained_name, **kwargs)
tokenizer_r = self.get_rust_tokenizer(pretrained_name, **kwargs)
tokenizer_p = self.get_tokenizer(pretrained_name, **kwargs)
ids_without_spe_char_r = tokenizer_r.encode(text_with_chinese_char, add_special_tokens=False)
ids_without_spe_char_p = tokenizer_p.encode(text_with_chinese_char, add_special_tokens=False)