add "ensure_variables_initialized" to Trainer

tfsnippet/trainer/base_trainer.py

@@ -79,14 +79,17 @@ class BaseTrainer(object):
         trainer.log_after_steps(1000)  # call `loop.print_logs` every 1000 steps
     """
 
-    def __init__(self, loop):
+    def __init__(self, loop, ensure_variables_initialized=True):
         """
         Initialize the internal states of :class:`BaseTrainer`.
 
         Args:
             loop (TrainLoop): The training loop object.
+            ensure_variables_initialized (bool): Whether or not to ensure
+                the variables are initialized in :meth:`run()`?
         """
         self._loop = loop
+        self._ensure_variables_initialized = ensure_variables_initialized
         self._events = EventSource([
             EventKeys.BEFORE_EXECUTION,
             EventKeys.AFTER_EXECUTION,
@@ -134,7 +137,8 @@ def run(self):
 
             # initialize global training status
             session = get_default_session_or_error()
-            ensure_variables_initialized()
+            if self._ensure_variables_initialized:
+                ensure_variables_initialized()
             self.loop.print_training_summary()
 
             for _ in self.loop.iter_epochs():

tfsnippet/trainer/trainer.py

@@ -65,7 +65,8 @@ class Trainer(BaseTrainer):
     """
 
     def __init__(self, loop, train_op, inputs, data_flow, feed_dict=None,
-                 metrics=None, summaries=None):
+                 metrics=None, summaries=None, 
+                 ensure_variables_initialized=True):
         """
 
         Args:
@@ -90,13 +91,18 @@ def __init__(self, loop, train_op, inputs, data_flow, feed_dict=None,
                 of summaries to be run and along with `train_op`, and later
                 to be added to ``loop.summary_writer``.
                 If ``loop.summary_writer`` is None, then no summary will be run.
+            ensure_variables_initialized (bool): Whether or not to ensure
+                the variables are initialized in :meth:`run()`?
         """
         if loop.max_epoch is None and loop.max_step is None:
             raise ValueError('At least one of `max_epoch`, `max_step` should '
                              'be configured for `loop`.')
         if summaries is not None and is_tensor_object(summaries):
             summaries = [summaries]
-        super(Trainer, self).__init__(loop=loop)
+        super(Trainer, self).__init__(
+            loop=loop,
+            ensure_variables_initialized=ensure_variables_initialized
+        )
 
         # memorize the arguments
         self._inputs = tuple(inputs or ())