Patch BigBird tokenization test (#12653)
This commit is contained in:
@@ -220,7 +220,7 @@ class BigBirdTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
|
|||||||
tokenizer = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base")
|
tokenizer = BigBirdTokenizer.from_pretrained("google/bigbird-roberta-base")
|
||||||
decoded_text = tokenizer.decode(tokenizer("Paris is the [MASK].").input_ids)
|
decoded_text = tokenizer.decode(tokenizer("Paris is the [MASK].").input_ids)
|
||||||
|
|
||||||
self.assertTrue(decoded_text == "[CLS] Paris is the [MASK].[SEP]")
|
self.assertTrue(decoded_text == "[CLS] Paris is the[MASK].[SEP]")
|
||||||
|
|
||||||
@slow
|
@slow
|
||||||
def test_tokenizer_integration(self):
|
def test_tokenizer_integration(self):
|
||||||
|
|||||||
Reference in New Issue
Block a user