make more callbacks triggerable

tensorpack/callbacks/base.py

@@ -3,11 +3,11 @@
 # Author: Yuxin Wu <ppwwyyxx@gmail.com>
 
 import tensorflow as tf
-from abc import ABCMeta
+from abc import ABCMeta, abstractmethod
 import six
 from ..tfutils.common import get_op_or_tensor_by_name, get_global_step_value
 
-__all__ = ['Callback', 'PeriodicCallback', 'ProxyCallback', 'CallbackFactory']
+__all__ = ['Callback', 'ProxyCallback', 'CallbackFactory', 'Triggerable']
 
 
 @six.add_metaclass(ABCMeta)
@@ -128,6 +128,39 @@ class Callback(object):
         return type(self).__name__
 
 
+@six.add_metaclass(ABCMeta)
+class Triggerable(Callback):
+    """
+    Base class for "triggerable" callback. It has a method :meth:`Triggerable.trigger()`
+    which can be triggered either inside an epoch or between epochs.
+    The higher-level wrapper will take the responsibility to determine when
+    to trigger.
+
+    If an triggerable is used as a callback directly (instead of under other
+    higher-level wrapper to control the trigger), it will by default trigger after
+    every epoch. This is mainly for backward-compatibilty and convenience.
+    """
+
+    def trigger(self):
+        """
+        Trigger something.
+        Note that this method may be called both inside an epoch and after an epoch.
+
+        Some operations (e.g. writing scalar stats) currently will cause
+        problems if run inside an epoch. This will be fixed in the future.
+        """
+        # TODO
+        self._trigger()
+
+    @abstractmethod
+    def _trigger(self):
+        pass
+
+    def _trigger_epoch(self):
+        """ If used as a callback directly, run the trigger every epoch."""
+        self.trigger()
+
+
 class ProxyCallback(Callback):
     """ A callback which proxy all methods to another callback.
         It's useful as a base class of callbacks which decorate other callbacks.
@@ -160,34 +193,6 @@ class ProxyCallback(Callback):
         return "Proxy-" + str(self.cb)
 
 
-class PeriodicCallback(ProxyCallback):
-    """
-    Wrap a callback so that after every ``period`` epochs, its :meth:`trigger_epoch`
-    method is called.
-    Note that this method will proxy the :meth:`trigger_step` method as-is.
-    """
-
-    def __init__(self, cb, period):
-        """
-        Args:
-            cb(Callback): the callback to be triggered periodically
-            period(int): the period, the number of epochs for a callback to be triggered.
-
-        Note:
-            In ``cb``, ``self.epoch_num`` will not be the true number of
-            epochs any more.
-        """
-        super(PeriodicCallback, self).__init__(cb)
-        self.period = int(period)
-
-    def _trigger_epoch(self):
-        if self.epoch_num % self.period == 0:
-            self.cb.trigger_epoch()
-
-    def __str__(self):
-        return "Periodic-" + str(self.cb)
-
-
 class CallbackFactory(Callback):
     """
     Create a callback with some lambdas.

tensorpack/callbacks/dump.py

@@ -6,7 +6,7 @@ import os
 import cv2
 import numpy as np
 
-from .trigger import Triggerable
+from .base import Triggerable
 from ..utils import logger
 from ..tfutils import get_op_tensor_name
 

tensorpack/callbacks/graph.py

@@ -5,7 +5,7 @@
 
 """ Graph related callbacks"""
 
-from .trigger import Triggerable
+from .base import Triggerable
 
 __all__ = ['RunOp']
 

tensorpack/callbacks/inference_runner.py

@@ -16,7 +16,7 @@ from ..tfutils import TowerContext
 from ..train.input_data import FeedfreeInput
 from ..predict import build_prediction_graph
 
-from .base import Callback
+from .base import Triggerable
 from .inference import Inferencer
 
 __all__ = ['InferenceRunner', 'FeedfreeInferenceRunner']
@@ -63,7 +63,7 @@ def summary_inferencer(trainer, infs):
             trainer.add_scalar_summary(k, v)
 
 
-class InferenceRunner(Callback):
+class InferenceRunner(Triggerable):
     """
     A callback that runs a list of :class:`Inferencer` on some
     :class:`DataFlow`.
@@ -128,7 +128,7 @@ class InferenceRunner(Callback):
         self.inf_to_tensors = [find_tensors(t) for t in dispatcer.get_names_for_each_entry()]
         # list of list of IOTensor
 
-    def _trigger_epoch(self):
+    def _trigger(self):
         for inf in self.infs:
             inf.before_inference()
 
@@ -147,7 +147,7 @@ class InferenceRunner(Callback):
         summary_inferencer(self.trainer, self.infs)
 
 
-class FeedfreeInferenceRunner(Callback):
+class FeedfreeInferenceRunner(Triggerable):
     """ A callback that runs a list of :class:`Inferencer` on some
     :class:`FeedfreeInput`, such as some tensor from a TensorFlow data reading
     pipeline.
@@ -231,7 +231,7 @@ class FeedfreeInferenceRunner(Callback):
         # list of list of id
         self.inf_to_idxs = dispatcer.get_idx_for_each_entry()
 
-    def _trigger_epoch(self):
+    def _trigger(self):
         for inf in self.infs:
             inf.before_inference()
 

tensorpack/callbacks/param.py

@@ -9,7 +9,7 @@ import operator
 import six
 import os
 
-from .trigger import Triggerable
+from .base import Triggerable
 from ..utils import logger
 from ..tfutils import get_op_tensor_name
 

tensorpack/callbacks/saver.py

@@ -6,10 +6,9 @@ import tensorflow as tf
 import os
 import shutil
 
-from .base import Callback
+from .base import Triggerable
 from ..utils import logger
 from ..tfutils.varmanip import get_savename_from_varname
-from .trigger import Triggerable
 
 __all__ = ['ModelSaver', 'MinSaver', 'MaxSaver']
 
@@ -83,7 +82,7 @@ due to an alternative in a different tower".format(v.name, var_dict[name].name))
             logger.exception("Exception in ModelSaver.trigger_epoch!")
 
 
-class MinSaver(Callback):
+class MinSaver(Triggerable):
     """
     Separately save the model with minimum value of some statistics.
     """
@@ -126,7 +125,7 @@ class MinSaver(Callback):
             return False
         return v > self.min if self.reverse else v < self.min
 
-    def _trigger_epoch(self):
+    def _trigger(self):
         if self.min is None or self._need_save():
             self.min = self._get_stat()
             if self.min:

tensorpack/callbacks/stats.py

@@ -6,8 +6,7 @@ import os
 import operator
 import json
 
-from .base import Callback
-from .trigger import Triggerable
+from .base import Triggerable
 from ..utils import logger
 
 __all__ = ['StatHolder', 'StatPrinter', 'SendStat']
@@ -110,7 +109,7 @@ class StatHolder(object):
             logger.exception("Exception in StatHolder.finalize()!")
 
 
-class StatPrinter(Callback):
+class StatPrinter(Triggerable):
     """
     A callback to control what stats to print. Enable by default to print
     everything in trainer.stat_holder.
@@ -132,7 +131,7 @@ class StatPrinter(Callback):
         # just try to add this stat earlier so SendStat can use
         self._stat_holder.add_stat('epoch_num', self.epoch_num + 1)
 
-    def _trigger_epoch(self):
+    def _trigger(self):
         # by default, add this two stat
         self._stat_holder.add_stat('global_step', self.global_step)
         self._stat_holder.finalize()

tensorpack/callbacks/trigger.py

@@ -3,46 +3,10 @@
 # File: trigger.py
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
-from abc import abstractmethod, ABCMeta
-import six
+from .base import ProxyCallback, Triggerable
 
-from .base import Callback, ProxyCallback
 
-
-__all__ = ['Triggerable', 'PeriodicTrigger']
-
-
-@six.add_metaclass(ABCMeta)
-class Triggerable(Callback):
-    """
-    Base class for "triggerable" callback. It has a method :meth:`Triggerable.trigger()`
-    which can be triggered either inside an epoch or between epochs.
-    The higher-level wrapper will take the responsibility to determine when
-    to trigger.
-
-    If an triggerable is used as a callback directly (instead of under other
-    higher-level wrapper to control the trigger), it will by default trigger after
-    every epoch. This is mainly for backward-compatibilty and convenience.
-    """
-
-    def trigger(self):
-        """
-        Trigger something.
-        Note that this method may be called both inside an epoch and after an epoch.
-
-        Some operations (e.g. writing scalar stats) currently will cause
-        problems if run inside an epoch. This will be fixed in the future.
-        """
-        # TODO
-        self._trigger()
-
-    @abstractmethod
-    def _trigger(self):
-        pass
-
-    def _trigger_epoch(self):
-        """ If used as a callback directly, run the trigger every epoch."""
-        self.trigger()
+__all__ = ['PeriodicTrigger', 'PeriodicCallback']
 
 
 class PeriodicTrigger(ProxyCallback):
@@ -78,3 +42,37 @@ class PeriodicTrigger(ProxyCallback):
             return
         if self.epoch_num % self._epoch_k == 0:
             self.cb.trigger()
+
+    def __str__(self):
+        return "PeriodicTrigger-" + str(self.cb)
+
+
+class PeriodicCallback(ProxyCallback):
+    """
+    Wrap a callback so that after every ``period`` epochs, its :meth:`trigger_epoch`
+    method is called.
+
+    Note that this wrapper will proxy the :meth:`trigger_step` method as-is.
+    To schedule a :class:`Triggerable` callback more frequent than once per
+    epoch, use :class:`PeriodicTrigger` instead.
+    """
+
+    def __init__(self, cb, period):
+        """
+        Args:
+            cb(Callback): the callback to be triggered periodically
+            period(int): the period, the number of epochs for a callback to be triggered.
+
+        Note:
+            In ``cb``, ``self.epoch_num`` will not be the true number of
+            epochs any more.
+        """
+        super(PeriodicCallback, self).__init__(cb)
+        self.period = int(period)
+
+    def _trigger_epoch(self):
+        if self.epoch_num % self.period == 0:
+            self.cb.trigger_epoch()
+
+    def __str__(self):
+        return "Periodic-" + str(self.cb)