add train

fluid/text_classification/train.py

@@ -85,9 +85,9 @@ def train_net():
             batch_size = 128, buf_size = 50000)
     
     if sys.argv[1] == "bow":
-        train(train_reader, word_dict, bow_net, use_cuda=False,
+        train(train_reader, word_dict, bow_net, use_cuda=True,
                 parallel=False, save_dirname="bow_model", lr=0.002,
-                pass_num=1, batch_size=128)
+                pass_num=30, batch_size=128)
     elif sys.argv[1] == "cnn":
         train(train_reader, word_dict, cnn_net, use_cuda=True,
                 parallel=False, save_dirname="cnn_model", lr=0.01,