Merge branch 'automodels' of https://github.com/huggingface/pytorch-transformers into automodels

pytorch_transformers/tokenization_transfo_xl.py

@@ -73,6 +73,10 @@ class TransfoXLTokenizer(PreTrainedTokenizer):
         super(TransfoXLTokenizer, self).__init__(unk_token=unk_token, eos_token=eos_token,
                                                  additional_special_tokens=additional_special_tokens,
                                                  **kwargs)
+
+        self.max_len_single_sentence = self.max_len # no default special tokens - you can update this value if you add special tokens
+        self.max_len_sentences_pair = self.max_len # no default special tokens - you can update this value if you add special tokens
+
         if never_split is None:
             never_split = self.all_special_tokens
         if special is None:

pytorch_transformers/tokenization_utils.py

@@ -166,6 +166,9 @@ class PreTrainedTokenizer(object):
         self._additional_special_tokens = []
 
         self.max_len = max_len if max_len is not None else int(1e12)
+        self.max_len_single_sentence = self.max_len
+        self.max_len_sentences_pair = self.max_len
+
         self.added_tokens_encoder = {}
         self.added_tokens_decoder = {}
 
@@ -590,10 +593,12 @@ class PreTrainedTokenizer(object):
             return first_sentence_tokens, second_sentence_tokens
 
     def add_special_tokens_single_sentence(self, token_ids):
-        raise NotImplementedError
+        logger.warning("This tokenizer does not make use of special tokens. The sequence has been returned with no modification.")
+        return token_ids
 
     def add_special_tokens_sentences_pair(self, token_ids_0, token_ids_1):
-        raise NotImplementedError
+        logger.warning("This tokenizer does not make use of special tokens. The two sequences have been concatenated.")
+        return token_ids_0 + token_ids_1
 
     def convert_ids_to_tokens(self, ids, skip_special_tokens=False):
         """ Converts a single index or a sequence of indices (integers) in a token "
@@ -636,9 +641,9 @@ class PreTrainedTokenizer(object):
         filtered_tokens = self.convert_ids_to_tokens(token_ids, skip_special_tokens=skip_special_tokens)
         text = self.convert_tokens_to_string(filtered_tokens)
 
-        if self.sep_token is not None and self.sep_token in text:
-            text = text.replace(self.cls_token, self.sep_token)
-            split_text = list(filter(lambda sentence: len(sentence) > 0, text.split(self.sep_token)))
+        if self._sep_token is not None and self._sep_token in text:
+            text = text.replace(self._cls_token, self._sep_token)
+            split_text = list(filter(lambda sentence: len(sentence) > 0, text.split(self._sep_token)))
             if clean_up_tokenization_spaces:
                 clean_text = [self.clean_up_tokenization(text) for text in split_text]
                 return clean_text

pytorch_transformers/tokenization_xlm.py

@@ -122,6 +122,10 @@ class XLMTokenizer(PreTrainedTokenizer):
                                            cls_token=cls_token, mask_token=mask_token,
                                            additional_special_tokens=additional_special_tokens,
                                            **kwargs)
+
+        self.max_len_single_sentence = self.max_len - 2  # take into account special tokens
+        self.max_len_sentences_pair = self.max_len - 3  # take into account special tokens
+
         try:
             import ftfy
             from spacy.lang.en import English

pytorch_transformers/tokenization_xlnet.py

@@ -71,6 +71,10 @@ class XLNetTokenizer(PreTrainedTokenizer):
                                              pad_token=pad_token, cls_token=cls_token,
                                              mask_token=mask_token, additional_special_tokens=
                                              additional_special_tokens, **kwargs)
+
+        self.max_len_single_sentence = self.max_len - 2  # take into account special tokens
+        self.max_len_sentences_pair = self.max_len - 3  # take into account special tokens
+
         try:
             import sentencepiece as spm
         except ImportError: