fix sentence piece input reader byte string (#453)

neulab · Jul 9, 2018 · 6fbbba4 · 6fbbba4
1 parent 753f246
commit 6fbbba4
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/xnmt/input_reader.py b/xnmt/input_reader.py
@@ -166,6 +166,7 @@ def read_sent(self, sentence):
       words = self.subword_model.SampleEncodeAsPieces(sentence.strip(), self.l, self.alpha)
     else:
       words = self.subword_model.EncodeAsPieces(sentence.strip())
+    words = [w.decode('utf-8') for w in words]
     return SimpleSentenceInput([self.vocab.convert(word) for word in words] + \
                                                        [self.vocab.convert(Vocab.ES_STR)], vocab_reference)