latest model link

tensorpack/callbacks/common.py

@@ -61,7 +61,17 @@ class ModelSaver(Callback):
                 self.path,
                 global_step=self.global_step,
                 write_meta_graph=False)
-        except Exception:   # disk error sometimes..
+
+            # create a symbolic link for the latest model
+            latest = self.saver.last_checkpoints[-1]
+            basename = os.path.basename(latest)
+            linkname = os.path.join(os.path.dirname(latest), 'latest')
+            try:
+                os.unlink(linkname)
+            except FileNotFoundError:
+                pass
+            os.symlink(basename, linkname)
+        except Exception:   # disk error sometimes.. just ignore
             logger.exception("Exception in ModelSaver.trigger_epoch!")
 
 class MinSaver(Callback):

tensorpack/callbacks/summary.py

@@ -22,6 +22,7 @@ class StatHolder(object):
         :param log_dir: directory to save the stats.
         """
         self.set_print_tag([])
+        self.blacklist_tag = set()
         self.stat_now = {}
 
         self.log_dir = log_dir
@@ -47,6 +48,9 @@ class StatHolder(object):
         """
         self.print_tag = None if print_tag is None else set(print_tag)
 
+    def add_blacklist_tag(self, blacklist_tag):
+        self.blacklist_tag |= set(blacklist_tag)
+
     def get_stat_now(self, key):
         """
         Return the value of a stat in the current epoch.
@@ -65,6 +69,7 @@ class StatHolder(object):
     def _print_stat(self):
         for k, v in sorted(self.stat_now.items(), key=operator.itemgetter(0)):
             if self.print_tag is None or k in self.print_tag:
+                if k not in self.blacklist_tag:
                     logger.info('{}: {:.5g}'.format(k, v))
 
     def _write_stat(self):

tensorpack/train/base.py

@@ -68,6 +68,7 @@ class Trainer(object):
         self._trigger_epoch()
         self.config.callbacks.trigger_epoch()
         self.summary_writer.flush()
+        self.stat_holder.add_stat('global_step', self.global_step)
         self.stat_holder.finalize()
 
     @abstractmethod
@@ -83,6 +84,8 @@ class Trainer(object):
         self.summary_op = tf.merge_all_summaries()
         # create an empty StatHolder
         self.stat_holder = StatHolder(logger.LOG_DIR)
+        # save global_step in stat.json, but don't print it
+        self.stat_holder.add_blacklist_tag(['global_step'])
 
     def _process_summary(self, summary_str):
         summary = tf.Summary.FromString(summary_str)