remove prefix from InputDesc.build_placeholder

tensorpack/graph_builder/model_desc.py

@@ -39,21 +39,17 @@ class InputDesc(
         self._cached_placeholder = None
         return self
 
-    def build_placeholder(self, prefix=''):
+    def build_placeholder(self):
         """
-        Build a tf.placeholder from the metadata, with an optional prefix.
-
-        Args:
-            prefix(str): the name of the placeholder will be ``prefix + self.name``
+        Build a tf.placeholder from the metadata.
 
         Returns:
             tf.Tensor:
         """
         with tf.name_scope(None):   # clear any name scope it might get called in
             ret = tf.placeholder(
-                self.type, shape=self.shape,
-                name=prefix + self.name)
-        if prefix == '' and self._cached_placeholder is None:
+                self.type, shape=self.shape, name=self.name)
+        if self._cached_placeholder is None:
             self._cached_placeholder = ret  # cached_placeholder only caches the prefix='' case
         return ret
 
@@ -97,8 +93,7 @@ class ModelDescBase(object):
         """
         Build the whole symbolic graph.
         This is supposed to be the "tower function" when used with :class:`TowerTrainer`.
-        By default it will call :meth:`_build_graph`
-        with a list of input tensors.
+        By default it will call :meth:`_build_graph` with a list of input tensors, for backward-compatibility.
 
         Args:
             args ([tf.Tensor]): tensors that matches the list of
@@ -108,10 +103,11 @@ class ModelDescBase(object):
             arg = args[0]
             if isinstance(arg, InputSource):
                 inputs = arg.get_input_tensors()  # remove in the future?
-                log_deprecated("build_graph(InputSource)", "Call with tensors in positional args instead.")
+                log_deprecated("build_graph(InputSource)",
+                               "Call with tensors in positional args instead.", "2018-03-31")
             elif isinstance(arg, (list, tuple)):
                 inputs = arg
-                log_deprecated("build_graph([Tensor])", "Call with positional args instead.")
+                log_deprecated("build_graph([Tensor])", "Call with positional args instead.", "2018-03-31")
             else:
                 inputs = [arg]
         else:

tensorpack/input_source/input_source.py

@@ -39,15 +39,8 @@ class PlaceholderInput(InputSource):
     """
     Just produce placeholders as input tensors.
     """
-    def __init__(self, prefix=''):
-        """
-        Args:
-            prefix(str): an optional prefix to add to the placeholder.
-        """
-        self._prefix = prefix
-
     def _setup(self, inputs):
-        self._all_placehdrs = [v.build_placeholder(prefix=self._prefix) for v in inputs]
+        self._all_placehdrs = [v.build_placeholder() for v in inputs]
 
     def _get_input_tensors(self):
         return self._all_placehdrs

tensorpack/predict/multigpu.py

@@ -6,6 +6,7 @@
 import tensorflow as tf
 from ..utils import logger
 from ..graph_builder.predict import SimplePredictBuilder
+from ..graph_builder.model_desc import InputDesc
 from ..input_source import PlaceholderInput
 from .base import OnlinePredictor
 
@@ -93,8 +94,11 @@ class DataParallelOfflinePredictor(OnlinePredictor):
 
             for idx, t in enumerate(towers):
                 tower_name = 'tower' + str(t)
-                input = PlaceholderInput(tower_name + '/')
-                input.setup(config.inputs_desc)
+
+                inputs_desc = [InputDesc(desc.type, desc.shape, tower_name + '/' + desc.name)
+                               for desc in config.inputs_desc]
+                input = PlaceholderInput()
+                input.setup(inputs_desc)
 
                 with tf.variable_scope(tf.get_variable_scope(), reuse=idx > 0):
                     builder = SimplePredictBuilder(ns_name=tower_name, device=t)

tensorpack/train/base.py

@@ -93,9 +93,11 @@ class Trainer(object):
     """ Base class for a trainer.
     """
 
-    _API_VERSION = 2
-
     is_chief = True
+    """
+    Whether this process is the chief worker in distributed training.
+    Certain callbacks will only be run by chief worker.
+    """
 
     def __init__(self, config=None):
         """

tensorpack/trainv1/base.py

@@ -40,8 +40,6 @@ class Trainer(object):
         monitors (Monitors): the monitors. Other callbacks can use it for logging.
     """
 
-    _API_VERSION = 1
-
     is_chief = True
     """
     Whether this process is the chief worker in distributed training.