make docstrings slightly more consistent

edward/inferences/gan_inference.py

@@ -62,7 +62,8 @@ def __init__(self, data, discriminator):
   def initialize(self, optimizer=None, optimizer_d=None,
                  global_step=None, global_step_d=None, var_list=None,
                  *args, **kwargs):
-    """Initialize inference algorithm.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       optimizer: str or tf.train.Optimizer, optional.

edward/inferences/gibbs.py

@@ -47,7 +47,8 @@ def __init__(self, latent_vars, proposal_vars=None, data=None):
     super(Gibbs, self).__init__(latent_vars, data)
 
   def initialize(self, scan_order='random', *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       scan_order: list or str, optional.

edward/inferences/hmc.py

@@ -49,7 +49,8 @@ def __init__(self, *args, **kwargs):
     super(HMC, self).__init__(*args, **kwargs)
 
   def initialize(self, step_size=0.25, n_steps=2, *args, **kwargs):
-    """Initialize inference algorithm.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       step_size: float, optional.

edward/inferences/implicit_klqp.py

@@ -81,7 +81,8 @@ def __init__(self, latent_vars, data=None, discriminator=None,
     super(GANInference, self).__init__(latent_vars, data)
 
   def initialize(self, ratio_loss='log', *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       ratio_loss: str or fn, optional.

edward/inferences/klpq.py

@@ -45,7 +45,8 @@ def __init__(self, *args, **kwargs):
     super(KLpq, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.

edward/inferences/klqp.py

@@ -51,7 +51,8 @@ def __init__(self, *args, **kwargs):
     super(KLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, kl_scaling=None, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -136,7 +137,8 @@ def __init__(self, *args, **kwargs):
     super(ReparameterizationKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -162,7 +164,8 @@ def __init__(self, *args, **kwargs):
     super(ReparameterizationKLKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, kl_scaling=None, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -202,7 +205,8 @@ def __init__(self, *args, **kwargs):
     super(ReparameterizationEntropyKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -229,7 +233,8 @@ def __init__(self, *args, **kwargs):
     super(ScoreKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -255,7 +260,8 @@ def __init__(self, *args, **kwargs):
     super(ScoreKLKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, kl_scaling=None, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.
@@ -295,7 +301,8 @@ def __init__(self, *args, **kwargs):
     super(ScoreEntropyKLqp, self).__init__(*args, **kwargs)
 
   def initialize(self, n_samples=1, *args, **kwargs):
-    """Initialization.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       n_samples: int, optional.

edward/inferences/map.py

@@ -69,7 +69,7 @@ def __init__(self, latent_vars=None, data=None):
 
     Args:
       latent_vars: list of RandomVariable or
-                    dict of RandomVariable to RandomVariable.
+                   dict of RandomVariable to RandomVariable.
         Collection of random variables to perform inference on. If
         list, each random variable will be implictly optimized
         using a `PointMass` random variable that is defined

edward/inferences/variational_inference.py

@@ -33,7 +33,8 @@ def __init__(self, *args, **kwargs):
 
   def initialize(self, optimizer=None, var_list=None, use_prettytensor=False,
                  global_step=None, *args, **kwargs):
-    """Initialize variational inference.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       optimizer: str or tf.train.Optimizer, optional.

edward/inferences/wgan_inference.py

@@ -22,6 +22,14 @@ class WGANInference(GANInference):
   models. These models do not require a tractable density and assume
   only a program that generates samples.
 
+  #### Notes
+
+  Argument-wise, the only difference from `GANInference` is
+  conceptual: the `discriminator` is better described as a test
+  function or critic. `WGANInference` continues to use
+  `discriminator` only to share methods and attributes with
+  `GANInference`.
+
   #### Examples
 
   ```python
@@ -30,20 +38,13 @@ class WGANInference(GANInference):
 
   inference = ed.WGANInference({x: x_data}, discriminator)
   ```
-
-  #### Notes
-
-  Argument-wise, the only difference from `GANInference` is
-  conceptual: the `discriminator` is better described as a test
-  function or critic. `WGANInference` continues to use
-  `discriminator` only to share methods and attributes with
-  `GANInference`.
   """
   def __init__(self, *args, **kwargs):
     super(WGANInference, self).__init__(*args, **kwargs)
 
   def initialize(self, penalty=10.0, clip=None, *args, **kwargs):
-    """Initialize inference algorithm.
+    """Initialize inference algorithm. It initializes hyperparameters
+    and builds ops for the algorithm's computation graph.
 
     Args:
       penalty: float, optional.