fixed decorator bug

antk/core/node_ops.py

@@ -293,7 +293,7 @@ def convolutional_net(in_progress=None):
 @neural_net
 def residual_dnn(tensor_in, hidden_units, activation='tanh', distribution='tnorm',
         initrange=1.0, l2=0.0, bn=False, keep_prob=None, fan_scaling=False,
-        skiplayers=3, name='dnn'):
+        skiplayers=3, name='residual_dnn'):
     """
     Creates residual neural network with shortcut connections.
         `Deep Residual Learning for Image Recognition`_
@@ -392,7 +392,7 @@ def highway_dnn(tensor_in, hidden_units, activation='tanh', distribution='tnorm'
     return tensor_in
 
 @node_op
-def dropout(tensor_in, prob, name=None):
+def dropout(tensor_in, prob, name='Dropout'):
     """
     Adds dropout node. Adapted from skflow `dropout_ops.py`_ .
         `Dropout A Simple Way to Prevent Neural Networks from Overfitting`_

antk/lib/decorate.py

@@ -36,8 +36,7 @@ def new_function(*args, **kwargs):
         if defaults:
             keyword_args = merge_dict(defaults, keyword_args)
         if 'name' in keyword_args:
-            with tf.variable_scope(keyword_args['name']):
-                tensor_out = func(*args, **kwargs)
+            tensor_out = func(*args, **kwargs)
 
             def node_repr(tensor_node):
                 return 'Tensor("%s", shape=%s, dtype=%r)' % (tensor_node.name,
@@ -75,11 +74,11 @@ def tanhlecun(tensor_in, name='tanhlecun'):
     return 1.7159*tf.nn.tanh((2.0/3.0) * tensor_in)
 
 
-sigmoid = act(tf.nn.sigmoid)
-tanh = act(tf.nn.tanh)
-relu = act(tf.nn.relu)
-relu6 = act(tf.nn.relu6)
-softplus = act(tf.nn.softplus)
+sigmoid = tf.nn.sigmoid
+tanh = tf.nn.tanh
+relu = tf.nn.relu
+relu6 = tf.nn.relu6
+softplus = tf.nn.softplus
 
 ACTIVATION = {'sigmoid': sigmoid,
               'tanh': tanh,