Merge pull request #1721 from huggingface/common_attributes

Add common getter and setter for input_embeddings & output_embeddings

transformers/tests/tokenization_xlnet_test.py

@@ -16,6 +16,7 @@ from __future__ import absolute_import, division, print_function, unicode_litera
 
 import os
 import unittest
+import pytest
 
 from transformers.tokenization_xlnet import (XLNetTokenizer, SPIECE_UNDERLINE)
 
@@ -89,6 +90,7 @@ class XLNetTokenizationTest(CommonTestCases.CommonTokenizerTester):
                                       u'9', u'2', u'0', u'0', u'0', u',', SPIECE_UNDERLINE + u'and', SPIECE_UNDERLINE + u'this',
                                       SPIECE_UNDERLINE + u'is', SPIECE_UNDERLINE + u'f', u'al', u'se', u'.'])
 
+    @pytest.mark.slow
     def test_sequence_builders(self):
         tokenizer = XLNetTokenizer.from_pretrained("xlnet-base-cased")