diff --git a/examples/keras_parikh_entailment/__main__.py b/examples/keras_parikh_entailment/__main__.py
index 927120f3c..5c3132bab 100644
--- a/examples/keras_parikh_entailment/__main__.py
+++ b/examples/keras_parikh_entailment/__main__.py
@@ -18,7 +18,7 @@ except ImportError:
     import pickle
 
 
-def train(model_dir, train_loc, dev_loc, shape, settings):
+def train(train_loc, dev_loc, shape, settings):
     train_texts1, train_texts2, train_labels = read_snli(train_loc)
     dev_texts1, dev_texts2, dev_labels = read_snli(dev_loc)
 
@@ -44,7 +44,7 @@ def train(model_dir, train_loc, dev_loc, shape, settings):
         batch_size=settings['batch_size'])
     if not (nlp.path / 'similarity').exists():
         (nlp.path / 'similarity').mkdir()
-    print("Saving to", model_dir / 'similarity')
+    print("Saving to", nlp.path / 'similarity')
     weights = model.get_weights()
     with (nlp.path / 'similarity' / 'model').open('wb') as file_:
         pickle.dump(weights[1:], file_)
@@ -68,8 +68,8 @@ def evaluate(model_dir, dev_loc):
     return correct, total
 
 
-def demo(model_dir):
-    nlp = spacy.load('en', path=model_dir,
+def demo():
+    nlp = spacy.load('en',
             create_pipeline=create_similarity_pipeline)
     doc1 = nlp(u'What were the best crime fiction books in 2016?')
     doc2 = nlp(
@@ -98,7 +98,6 @@ def read_snli(path):
 
 @plac.annotations(
     mode=("Mode to execute", "positional", None, str, ["train", "evaluate", "demo"]),
-    model_dir=("Path to spaCy model directory", "positional", None, Path),
     train_loc=("Path to training data", "positional", None, Path),
     dev_loc=("Path to development data", "positional", None, Path),
     max_length=("Length to truncate sentences", "option", "L", int),
@@ -110,7 +109,7 @@ def read_snli(path):
     tree_truncate=("Truncate sentences by tree distance", "flag", "T", bool),
     gru_encode=("Encode sentences with bidirectional GRU", "flag", "E", bool),
 )
-def main(mode, model_dir, train_loc, dev_loc,
+def main(mode, train_loc, dev_loc,
         tree_truncate=False,
         gru_encode=False,
         max_length=100,
@@ -129,12 +128,12 @@ def main(mode, model_dir, train_loc, dev_loc,
         'gru_encode': gru_encode
     }
     if mode == 'train':
-        train(model_dir, train_loc, dev_loc, shape, settings)
+        train(train_loc, dev_loc, shape, settings)
     elif mode == 'evaluate':
-        correct, total = evaluate(model_dir, dev_loc)
+        correct, total = evaluate(dev_loc)
         print(correct, '/', total, correct / total)
     else:
-        demo(model_dir)
+        demo()
 
 if __name__ == '__main__':
     plac.call(main)