diff --git a/BERT/reader/pretraining.py b/BERT/reader/pretraining.py
index 6f99d6e36d559d72e993f4e08f942c2168c3dda2..c21a43d33caedd9a01c02dacbedd01a16e1eec9f 100644
--- a/BERT/reader/pretraining.py
+++ b/BERT/reader/pretraining.py
@@ -78,7 +78,7 @@ class DataReader(object):
     def parse_line(self, line, max_seq_len=512):
         """ parse one line to token_ids, sentence_ids, pos_ids, label
         """
-        line = line.strip().split(";")
+        line = line.strip().decode().split(";")
         assert len(line) == 4, "One sample must have 4 fields!"
         (token_ids, sent_ids, pos_ids, label) = line
         token_ids = [int(token) for token in token_ids.split(" ")]