Merge pull request #293 from sony/feature/20181107-pf-batch-norm-init…

…ializers

 [python] Initialization option in batch normalization parametric function

python/src/nnabla/parametric_functions.py

@@ -1312,7 +1312,8 @@ def depthwise_deconvolution(inp, kernel, pad=None, stride=None, dilation=None,
     ('var', 'Moving average of batch variance', '<see above>', False),
 ])
 def batch_normalization(inp, axes=[1], decay_rate=0.9, eps=1e-5,
-                        batch_stat=True, output_stat=False, fix_parameters=False):
+                        batch_stat=True, output_stat=False, fix_parameters=False,
+                        param_init=None):
     """
     Batch normalization layer.
 
@@ -1341,6 +1342,12 @@ def batch_normalization(inp, axes=[1], decay_rate=0.9, eps=1e-5,
         batch_stat (bool): Use mini-batch statistics rather than running ones.
         output_stat (bool): Output batch mean and variance.
         fix_parameters (bool): When set to `True`, the beta and gamma will not be updated.
+        param_init (dict):
+            Parameter initializers can be set with a dict. A key of the dict must
+            be ``'beta'``, ``'gamma'``, ``'mean'`` or ``'var'``.
+            A value of the dict must be an :obj:`~nnabla.initializer.Initializer`
+            or a :obj:`numpy.ndarray`.
+            E.g. ``{'beta': ConstantIntializer(0), 'gamma': np.ones(gamma_shape) * 2}``.
 
     Returns:
         :class:`~nnabla.Variable`: N-D array.
@@ -1359,14 +1366,21 @@ def batch_normalization(inp, axes=[1], decay_rate=0.9, eps=1e-5,
     assert len(axes) == 1
     shape_stat = [1 for _ in inp.shape]
     shape_stat[axes[0]] = inp.shape[axes[0]]
+
+    if param_init is None:
+        param_init = {}
+    beta_init = param_init.get('beta', ConstantInitializer(0))
+    gamma_init = param_init.get('gamma', ConstantInitializer(1))
+    mean_init = param_init.get('mean', ConstantInitializer(0))
+    var_init = param_init.get('var', ConstantInitializer(0))
     beta = get_parameter_or_create(
-        "beta", shape_stat, ConstantInitializer(0), True, not fix_parameters)
+        "beta", shape_stat, beta_init, True, not fix_parameters)
     gamma = get_parameter_or_create(
-        "gamma", shape_stat, ConstantInitializer(1), True, not fix_parameters)
+        "gamma", shape_stat, gamma_init, True, not fix_parameters)
     mean = get_parameter_or_create(
-        "mean", shape_stat, ConstantInitializer(0), False)
+        "mean", shape_stat, mean_init, False)
     var = get_parameter_or_create(
-        "var", shape_stat, ConstantInitializer(0), False)
+        "var", shape_stat, var_init, False)
     return F.batch_normalization(inp, beta, gamma, mean, var, axes,
                                  decay_rate, eps, batch_stat, output_stat)