print summary

example_cifar10.py

@@ -134,7 +134,7 @@ def get_config():
         dataset=dataset_train,
         optimizer=tf.train.GradientDescentOptimizer(lr),
         callbacks=Callbacks([
-            SummaryWriter(),
+            SummaryWriter(print_tag=['train_cost', 'train_error']),
             PeriodicSaver(),
             ValidationError(dataset_test, prefix='test'),
         ]),

example_mnist.py

@@ -126,7 +126,7 @@ def get_config():
         dataset=dataset_train,
         optimizer=tf.train.AdamOptimizer(lr),
         callbacks=Callbacks([
-            SummaryWriter(),
+            SummaryWriter(print_tag=['train_cost', 'train_error']),
             PeriodicSaver(),
             ValidationError(dataset_test, prefix='test'),
         ]),

tensorpack/callbacks/common.py

@@ -30,8 +30,9 @@ class PeriodicSaver(PeriodicCallback):
             global_step=self.global_step)
 
 class SummaryWriter(Callback):
-    def __init__(self):
+    def __init__(self, print_tag=None):
         self.log_dir = logger.LOG_DIR
+        self.print_tag = print_tag if print_tag else ['train_cost']
 
     def _before_train(self):
         self.writer = tf.train.SummaryWriter(
@@ -44,5 +45,11 @@ class SummaryWriter(Callback):
         if self.summary_op is None:
             return
         summary_str = self.summary_op.eval()
-        self.writer.add_summary(summary_str, get_global_step())
+        summary = tf.Summary.FromString(summary_str)
+        for val in summary.value:
+            if val.tag in self.print_tag:
+                assert val.WhichOneof('value') == 'simple_value', \
+                    'Cannot print summary {}: not a simple_value summary!'.format(val.tag)
+                logger.info('{}: {:.4f}'.format(val.tag, val.simple_value))
+        self.writer.add_summary(summary, get_global_step())
 

tensorpack/callbacks/group.py

@@ -58,9 +58,9 @@ class CallbackTimeLogger(object):
         msgs = []
         for name, t in self.times:
             if t / self.tot > 0.3 and t > 1:
-                msgs.append("{}:{}sec".format(name, t))
+                msgs.append("{}:{:.3f}sec".format(name, t))
         logger.info(
-            "Callbacks took {} sec in total. {}".format(
+            "Callbacks took {:.3f} sec in total. {}".format(
                 self.tot, ' '.join(msgs)))
 
 

tensorpack/callbacks/validation_callback.py

@@ -60,14 +60,9 @@ class ValidationError(PeriodicCallback):
                 pbar.update()
 
         cost_avg = cost_sum / cnt
-        self.writer.add_summary(
-            create_summary('{}_error'.format(self.prefix),
-                           err_stat.accuracy),
-            self.global_step)
-        self.writer.add_summary(
-            create_summary('{}_cost'.format(self.prefix),
-                           cost_avg),
-            self.global_step)
-        logger.info(
-            "{} validation after epoch{},step{}: err={:.4f}, cost={:.3f}".format(
-            self.prefix, self.epoch_num, self.global_step, err_stat.accuracy, cost_avg))
+        self.writer.add_summary(create_summary(
+            '{}_error'.format(self.prefix), err_stat.accuracy), self.global_step)
+        self.writer.add_summary(create_summary(
+            '{}_cost'.format(self.prefix), cost_avg), self.global_step)
+        logger.info("{}_cost: {:.4f}".format(self.prefix, cost_avg))
+        logger.info("{}_error: {:.4f}".format(self.prefix, err_stat.accuracy))

tensorpack/utils/summary.py

@@ -56,7 +56,10 @@ def summary_moving_average(cost_var):
             tf.get_collection(SUMMARY_VARS_KEY) + \
             tf.get_collection(COST_VARS_KEY)
     avg_maintain_op = averager.apply(vars_to_summary)
-    for c in vars_to_summary:
-        tf.scalar_summary(c.op.name, averager.average(c))
+    for idx, c in enumerate(vars_to_summary):
+        name = c.op.name
+        if idx == 0:
+            name = 'train_cost'
+        tf.scalar_summary(name, averager.average(c))
     return avg_maintain_op