update docs. remove _register_monitor

tensorpack/graph_builder/model_desc.py

@@ -67,6 +67,11 @@ class InputDesc(
             return self._cached_placeholder
         return self.build_placeholder()
 
+    @staticmethod
+    def from_tensor(t):
+        return InputDesc(
+            t.dtype, t.shape.as_list(), t.name[:-2])
+
 
 @six.add_metaclass(ABCMeta)
 class ModelDescBase(object):

tensorpack/train/base.py

@@ -76,13 +76,13 @@ class Trainer(object):
         """
         self._callbacks = []
         self.loop = TrainLoop()
-        self._monitors = []  # Clarify the type. Don't change from list to monitors.
 
         # Hacks!
         if config is not None:
             logger.warn("You're initializing new trainer with old trainer API!")
             logger.warn("This could happen if you wrote a custom trainer before.")
             logger.warn("It may work now through some hacks, but please switch to the new API!")
+            logger.warn("See https://github.com/ppwwyyxx/tensorpack/issues/458 for more information.")
             self._config = config
             self.inputs_desc = config.model.get_inputs_desc()
             self.tower_func = TowerFuncWrapper(
@@ -117,19 +117,6 @@ class Trainer(object):
         else:
             self._callbacks.append(cb)
 
-    def _register_monitor(self, mon):
-        """
-        Register a monitor to the trainer.
-        It can only be called before :meth:`Trainer.train` gets called.
-        """
-        assert isinstance(mon, TrainingMonitor), mon
-        assert not isinstance(self._monitors, Monitors), \
-            "Cannot register more monitors after trainer was setup!"
-        if not self.is_chief and mon.chief_only:
-            logger.warn("Monitor {} is chief-only, skipped.".format(str(mon)))
-        else:
-            self._register_callback(mon)
-
     def run_step(self):
         """
         Defines what to do in one iteration. The default is:
@@ -154,8 +141,10 @@ class Trainer(object):
         self._register_callback(MaintainStepCounter())
         for cb in callbacks:
             self._register_callback(cb)
+        for cb in self._callbacks:
+            assert not isinstance(cb, TrainingMonitor), "Monitor cannot be pre-registered for now!"
         for m in monitors:
-            self._register_monitor(m)
+            self._register_callback(m)
         self.monitors = Monitors(monitors)
         self._register_callback(self.monitors)   # monitors is also a callback
 
@@ -283,8 +272,7 @@ class Trainer(object):
             else:
                 logger.warn("You're calling new trainers with old trainer API!")
                 logger.warn("Now it returns the old trainer for you, please switch to use new trainers soon!")
-                logger.warn("'SomeTrainer(config, ...).train()' should be equivalent to "
-                            "'launch_train_with_config(config, SomeTrainer(...))' in the new API.")
+                logger.warn("See https://github.com/ppwwyyxx/tensorpack/issues/458 for more information.")
                 return old_trainer(*args, **kwargs)
         else:
             return super(Trainer, cls).__new__(cls)

tensorpack/train/tower.py

@@ -123,13 +123,13 @@ class SingleCostTrainer(TowerTrainer):
         get_opt_fn = memoized(get_opt_fn)
         self.set_tower_func(get_cost_fn)
 
+        # TODO setup may want to register monitor as well??
         input_callbacks = self._setup_input(inputs_desc, input)
         train_callbacks = self._setup_graph(input, get_cost_fn, get_opt_fn)
         internal_callbacks = input_callbacks + train_callbacks
         for cb in internal_callbacks:
             self._register_callback(cb)
 
-    # TODO register directly instead of return?
     @abstractmethod
     def _setup_graph(self, input, get_cost_fn, get_opt_fn):
         """