remove Trainer._process_summary and use a general Trainer.add_summary

examples/DeepQNetwork/common.py

@@ -114,5 +114,5 @@ class Evaluator(Callback):
         t = time.time() - t
         if t > 10 * 60:  # eval takes too long
             self.eval_episode = int(self.eval_episode * 0.94)
-        self.trainer.write_scalar_summary('mean_score', mean)
-        self.trainer.write_scalar_summary('max_score', max)
+        self.trainer.add_scalar_summary('mean_score', mean)
+        self.trainer.add_scalar_summary('max_score', max)

examples/PennTreebank/PTB-LSTM.py

@@ -135,7 +135,7 @@ def get_config():
             FeedfreeInferenceRunner(val_data, [ScalarStats(['cost'])]),
             CallbackFactory(
                 trigger_epoch=lambda self:
-                self.trainer.write_scalar_summary(
+                self.trainer.add_scalar_summary(
                     'validation_perplexity',
                     np.exp(self.trainer.stat_holder.get_stat_now('validation_cost') / SEQ_LEN))),
             RunOp(lambda: M.reset_lstm_state()),

tensorpack/RL/expreplay.py

@@ -207,8 +207,8 @@ class ExpReplay(DataFlow, Callback):
         for k, v in six.iteritems(stats):
             try:
                 mean, max = np.mean(v), np.max(v)
-                self.trainer.write_scalar_summary('expreplay/mean_' + k, mean)
-                self.trainer.write_scalar_summary('expreplay/max_' + k, max)
+                self.trainer.add_scalar_summary('expreplay/mean_' + k, mean)
+                self.trainer.add_scalar_summary('expreplay/max_' + k, max)
             except:
                 pass
         self.player.reset_stat()

tensorpack/callbacks/inference_runner.py

@@ -58,7 +58,7 @@ def summary_inferencer(trainer, infs):
             except:
                 logger.warn("{} returns a non-scalar statistics!".format(type(inf).__name__))
                 continue
-            trainer.write_scalar_summary(k, v)
+            trainer.add_scalar_summary(k, v)
 
 
 class InferenceRunner(Callback):

tensorpack/tfutils/summary.py

@@ -13,11 +13,11 @@ from .tower import get_current_tower_context
 from . import get_global_step_var
 from .symbolic_functions import rms
 
-__all__ = ['create_summary', 'add_param_summary', 'add_activation_summary',
+__all__ = ['create_scalar_summary', 'add_param_summary', 'add_activation_summary',
            'add_moving_summary', 'summary_moving_average']
 
 
-def create_summary(name, v):
+def create_scalar_summary(name, v):
     """
     Returns:
         tf.Summary: a tf.Summary object with name and simple scalar value v.

tensorpack/train/base.py

@@ -16,7 +16,7 @@ from ..utils.timer import timed_operation
 from ..callbacks import StatHolder
 from ..tfutils import get_global_step, get_global_step_var
 from ..tfutils.modelutils import describe_model
-from ..tfutils.summary import create_summary
+from ..tfutils.summary import create_scalar_summary
 
 __all__ = ['Trainer', 'StopTraining']
 
@@ -96,8 +96,14 @@ class Trainer(object):
     def _trigger_epoch(self):
         pass
 
-    def _process_summary(self, summary_str):
-        summary = tf.Summary.FromString(summary_str)
+    def add_summary(self, summary):
+        """
+        Add summary to ``self.summary_writer``, and also
+        add scalar summary to ``self.stat_holder``.
+
+        Args:
+            summary (tf.Summary): a summary object.
+        """
         for val in summary.value:
             if val.WhichOneof('value') == 'simple_value':
                 val.tag = re.sub('tower[p0-9]+/', '', val.tag)   # TODO move to subclasses
@@ -107,17 +113,15 @@ class Trainer(object):
                 self.stat_holder.add_stat(val.tag, val.simple_value)
         self.summary_writer.add_summary(summary, get_global_step())
 
-    def write_scalar_summary(self, name, val):
+    def add_scalar_summary(self, name, val):
         """
-        Write a scalar sumary to both TF events file and StatHolder.
+        Add a scalar sumary to both TF events file and StatHolder.
 
         Args:
             name(str)
             val(float)
         """
-        self.summary_writer.add_summary(
-            create_summary(name, val), get_global_step())
-        self.stat_holder.add_stat(name, val)
+        self.add_summary(create_scalar_summary(name, val))
 
     def setup(self):
         """

tensorpack/train/feedfree.py

@@ -29,7 +29,7 @@ class FeedfreeTrainerBase(Trainer):
         # note that summary_op will take a data from the queue
         if self.summary_op is not None:
             summary_str = self.summary_op.eval()
-            self._process_summary(summary_str)
+            self.add_summary(tf.Summary.FromString(summary_str))
 
     def _get_input_tensors(self):
         return self._input_method.get_input_tensors()

tensorpack/train/multigpu.py

@@ -209,7 +209,7 @@ class AsyncMultiGPUTrainer(MultiGPUTrainer,
             if self.config.tower > 1:
                 async_step_total_cnt = int(re.findall(
                     '[0-9]+', self.async_step_counter.__str__())[0])
-                self.write_scalar_summary(
+                self.add_scalar_summary(
                     'async_global_step', async_step_total_cnt)
         except:
             logger.exception("Cannot log async_global_step")

tensorpack/train/trainer.py

@@ -95,7 +95,7 @@ class SimpleTrainer(Trainer):
         if self.summary_op is not None:
             feed = self._input_method.next_feed()
             summary_str = self.summary_op.eval(feed_dict=feed)
-            self._process_summary(summary_str)
+            self.add_summary(tf.Summary.FromString(summary_str))
 
     def get_predict_func(self, input_names, output_names):
         return self._predictor_factory.get_predictor(input_names, output_names, 0)