use "tensor" instead of "var"

tensorpack/callbacks/param.py

@@ -11,7 +11,7 @@ import os
 
 from .base import Callback
 from ..utils import logger
-from ..tfutils import get_op_var_name
+from ..tfutils import get_op_tensor_name
 
 __all__ = ['HyperParam', 'GraphVarParam', 'ObjAttrParam',
            'HyperParamSetter', 'HumanHyperParamSetter',
@@ -62,7 +62,7 @@ class GraphVarParam(HyperParam):
         """
         self.name = name
         self.shape = shape
-        self._readable_name, self.var_name = get_op_var_name(name)
+        self._readable_name, self.var_name = get_op_tensor_name(name)
 
     def setup_graph(self):
         """ Will setup the assign operator for that variable. """

tensorpack/tfutils/__init__.py

@@ -17,8 +17,8 @@ def _global_import(name):
 
 
 _TO_IMPORT = set([
-    'sessinit',
     'common',
+    'sessinit',
     'gradproc',
     'argscope',
     'tower'

tensorpack/tfutils/common.py

@@ -12,10 +12,9 @@ from contextlib import contextmanager
 __all__ = ['get_default_sess_config',
            'get_global_step',
            'get_global_step_var',
-           'get_op_var_name',
            'get_op_tensor_name',
-           'get_vars_by_names',
            'get_tensors_by_names',
+           'get_op_or_tensor_by_name',
            'backup_collection',
            'restore_collection',
            'clear_collection',
@@ -87,9 +86,6 @@ def get_op_tensor_name(name):
         return name, name + ':0'
 
 
-get_op_var_name = get_op_tensor_name
-
-
 def get_tensors_by_names(names):
     """
     Get a list of tensors in the default graph by a list of names.
@@ -100,12 +96,17 @@ def get_tensors_by_names(names):
     ret = []
     G = tf.get_default_graph()
     for n in names:
-        opn, varn = get_op_var_name(n)
+        opn, varn = get_op_tensor_name(n)
         ret.append(G.get_tensor_by_name(varn))
     return ret
 
 
-get_vars_by_names = get_tensors_by_names
+def get_op_or_tensor_by_name(name):
+    G = tf.get_default_graph()
+    if len(name) >= 3 and name[-2] == ':':
+        return G.get_tensor_by_name(name)
+    else:
+        return G.get_operation_by_name(name)
 
 
 def backup_collection(keys):

tensorpack/tfutils/sessinit.py

@@ -10,7 +10,7 @@ import tensorflow as tf
 import six
 
 from ..utils import logger, PREDICT_TOWER
-from .common import get_op_var_name
+from .common import get_op_tensor_name
 from .varmanip import (SessionUpdate, get_savename_from_varname,
                        is_training_name, get_checkpoint_path)
 
@@ -149,7 +149,7 @@ class ParamRestore(SessionInit):
             param_dict (dict): a dict of {name: value}
         """
         # use varname (with :0) for consistency
-        self.prms = {get_op_var_name(n)[1]: v for n, v in six.iteritems(param_dict)}
+        self.prms = {get_op_tensor_name(n)[1]: v for n, v in six.iteritems(param_dict)}
 
     def _init(self, sess):
         variables = tf.get_collection(tf.GraphKeys().VARIABLES)  # TODO