From c6acbdd50af6d1b915916001c01c6d2760fe316d Mon Sep 17 00:00:00 2001
From: Julien Chaumond <chaumond@gmail.com>
Date: Fri, 27 Sep 2019 17:02:29 -0400
Subject: [PATCH] Close #1304

---
 transformers/tokenization_roberta.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/transformers/tokenization_roberta.py b/transformers/tokenization_roberta.py
index 32e92be211..7adeea689e 100644
--- a/transformers/tokenization_roberta.py
+++ b/transformers/tokenization_roberta.py
@@ -81,6 +81,8 @@ class RobertaTokenizer(GPT2Tokenizer):
                                                bos_token=bos_token, eos_token=eos_token, unk_token=unk_token,
                                                sep_token=sep_token, cls_token=cls_token, pad_token=pad_token,
                                                mask_token=mask_token, **kwargs)
+        self.max_len_single_sentence = self.max_len - 2  # take into account special tokens
+        self.max_len_sentences_pair = self.max_len - 4  # take into account special tokens
 
     def add_special_tokens_single_sequence(self, token_ids):
         """