diff --git a/examples/chainer_sentiment.py b/examples/chainer_sentiment.py
index 251eadef3..ac3881e75 100644
--- a/examples/chainer_sentiment.py
+++ b/examples/chainer_sentiment.py
@@ -64,6 +64,7 @@ class SentimentAnalyser(object):
         # For arbitrary data storage, there's:
         # doc.user_data['my_data'] = y
 
+
 class Classifier(Chain):
     def __init__(self, predictor):
         super(Classifier, self).__init__(predictor=predictor)
@@ -77,9 +78,10 @@ class Classifier(Chain):
 
 
 class SentimentModel(Chain):
-    def __init__(self, shape, **settings):
+    def __init__(self, nlp, shape, **settings):
         Chain.__init__(self,
-            embed=_Embed(shape['nr_vector'], shape['nr_dim'], shape['nr_hidden']),
+            embed=_Embed(shape['nr_vector'], shape['nr_dim'], shape['nr_hidden'],
+                initialW=lambda arr: set_vectors(arr, nlp.vocab)),
             encode=_Encode(shape['nr_hidden'], shape['nr_hidden']),
             attend=_Attend(shape['nr_hidden'], shape['nr_hidden']),
             predict=_Predict(shape['nr_hidden'], shape['nr_class']))
@@ -205,16 +207,14 @@ def get_features(docs, max_length):
     return Xs
 
 
-def get_embeddings(vocab, max_rank=1000):
-    if max_rank is None:
-        max_rank = max(lex.rank+1 for lex in vocab if lex.has_vector)
-    vectors = xp.ndarray((max_rank+1, vocab.vectors_length), dtype='f')
+def set_vectors(vectors, vocab):
     for lex in vocab:
-        if lex.has_vector and lex.rank < max_rank:
+        if lex.has_vector and (lex.rank+1) < vectors.shape[0]:
             lex.norm = lex.rank+1
             vectors[lex.rank + 1] = lex.vector
         else:
             lex.norm = 0
+    vectors.unchain_backwards()
     return vectors
 
 
@@ -222,13 +222,10 @@ def train(train_texts, train_labels, dev_texts, dev_labels,
         lstm_shape, lstm_settings, lstm_optimizer, batch_size=100, nb_epoch=5,
         by_sentence=True):
     nlp = spacy.load('en', entity=False)
-    for lex in nlp.vocab:
-        if lex.rank >= (lstm_shape['nr_vector'] - 1):
-            lex.norm = 0
-        else:
-            lex.norm = lex.rank+1
+    if 'nr_vector' not in lstm_shape:
+        lstm_shape['nr_vector'] = max(lex.rank+1 for lex in vocab if lex.has_vector)
     print("Make model")
-    model = Classifier(SentimentModel(lstm_shape, **lstm_settings))
+    model = Classifier(SentimentModel(nlp, lstm_shape, **lstm_settings))
     print("Parsing texts...")
     if by_sentence:
         train_data = SentenceDataset(nlp, train_texts, train_labels, lstm_shape['max_length'])