Merge pull request #684 from jmhessel/mergefixes

Added averaging support in merge and a TimeDistributedMerge layer

keras/layers/core.py

@@ -161,11 +161,42 @@ class Masking(MaskedLayer):
         return {"name": self.__class__.__name__,
                 "mask_value": self.mask_value}
 
+class TimeDistributedMerge(Layer):
+    def __init__(self, mode='sum'):
+        '''
+        Sum/multiply/avearge over a time distributed layer's outputs.
+        mode: {'sum', 'mul', 'ave'}
+        Tensor input dimensions:   (nb_sample, shared_dimension, input_dim)
+        Tensor output dimensions:  (nb_sample, output_dim)
+        '''
+        self.mode = mode
+        self.params = []
+        self.regularizers = []
+        self.constraints = []
+        self.updates = []
+
+    def get_output(self, train=False):
+        X = self.get_input(train)
+        if self.mode == 'sum' or self.mode == 'ave':
+            s = theano.tensor.sum(X, axis=1)
+            if self.mode == 'ave':
+                s /= X.shape[1]
+            return s
+        elif self.mode == 'mul':
+            s = theano.tensor.mul(X, axis=1)
+            return s
+        else:
+            raise Exception('Unknown merge mode')
+
+    def get_config(self):
+        return {"name": self.__class__.__name__,
+                "mode": self.mode}
+
 
 class Merge(Layer):
     def __init__(self, layers, mode='sum', concat_axis=-1):
         ''' Merge the output of a list of layers or containers into a single tensor.
-            mode: {'sum', 'mul', 'concat'}
+            mode: {'sum', 'mul', 'concat', 'ave'}
         '''
         if len(layers) < 2:
             raise Exception("Please specify two or more input layers (or containers) to merge")
@@ -190,10 +221,12 @@ class Merge(Layer):
         return self.params, self.regularizers, self.constraints, self.updates
 
     def get_output(self, train=False):
-        if self.mode == 'sum':
+        if self.mode == 'sum' or self.mode == 'ave':
             s = self.layers[0].get_output(train)
             for i in range(1, len(self.layers)):
                 s += self.layers[i].get_output(train)
+            if self.mode == 'ave':
+                s /= len(self.layers)
             return s
         elif self.mode == 'concat':
             inputs = [self.layers[i].get_output(train) for i in range(len(self.layers))]

tests/auto/keras/layers/test_core.py

@@ -104,6 +104,10 @@ class TestConfigParams(unittest.TestCase):
         layer = core.TimeDistributedDense(10, 10)
         self._runner(layer)
 
+    def test_time_dist_merge(self):
+        layer = core.TimeDistributedMerge()
+        self._runner(layer)
+
     def test_autoencoder(self):
         layer_1 = core.Layer()
         layer_2 = core.Layer()