Merge pull request yunjey#30 from keishinkickback/master

Bug fixes in /tutorials/09 - Image Captioning/train.py

Author: yunjey

tutorials/09 - Image Captioning/train.py

@@ -3,6 +3,7 @@
 import torch.nn as nn
 import numpy as np
 import os
+import pickle
 from data_loader import get_loader 
 from build_vocab import Vocabulary
 from model import EncoderCNN, DecoderRNN 
@@ -24,7 +25,7 @@ def main(args):
         transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
 
     # Load vocabulary wrapper.
-    with open(vocab_path, 'rb') as f:
+    with open(args.vocab_path, 'rb') as f:
         vocab = pickle.load(f)
 
     # Build data loader
@@ -115,4 +116,3 @@ def main(args):
     parser.add_argument('--learning_rate', type=float, default=0.001)
     args = parser.parse_args()
     print(args)
-    main(args)