diff --git a/rasa_core/train.py b/rasa_core/train.py
index f7f9e1c72d3..a18aaff7390 100644
--- a/rasa_core/train.py
+++ b/rasa_core/train.py
@@ -5,9 +5,12 @@
 
 import argparse
 import logging
+
 from builtins import str
 
 from rasa_core.agent import Agent
+from rasa_core.channels.console import ConsoleInputChannel
+from rasa_core.interpreter import RasaNLUInterpreter, RegexInterpreter
 from rasa_core.policies.keras_policy import KerasPolicy
 from rasa_core.policies.memoization import MemoizationPolicy
 
@@ -54,8 +57,7 @@ def create_argument_parser():
     parser.add_argument(
             '--online',
             default=False,
-            action='store_const',
-            const=True,
+            action='store_true',
             help="enable online training")
     parser.add_argument(
             '--augmentation',
@@ -66,16 +68,13 @@ def create_argument_parser():
 
 
 def train_dialogue_model(domain_file, stories_file, output_path,
-                         online, nlu_model_path, kwargs):
+                         use_online_learning, nlu_model_path, kwargs):
     agent = Agent(domain_file, policies=[MemoizationPolicy(), KerasPolicy()])
 
-    if online:
-        from rasa_core.channels.console import ConsoleInputChannel
-        from rasa_core.interpreter import RasaNLUInterpreter
+    if use_online_learning:
         if nlu_model_path:
             agent.interpreter = RasaNLUInterpreter(nlu_model_path)
         else:
-            from rasa_core.interpreter import RegexInterpreter
             agent.interpreter = RegexInterpreter()
         agent.train_online(
                 stories_file,