Test for various layer types including Gaussian dropout and leaky rec…

…tifier.

sknn/backend/lasagne/mlp.py

@@ -145,6 +145,8 @@ def _create_convolution_layer(self, name, layer, network):
         return network
 
     def _create_native_layer(self, name, layer, network):
+        if layer.units and 'num_units' not in layer.keywords:
+            layer.keywords['num_units'] = layer.units
         return layer.type(network, *layer.args, **layer.keywords)
 
     def _create_layer(self, name, layer, network):

sknn/nn.py

@@ -29,9 +29,9 @@ class ansi:
 
 class Native(object):
 
-    def __init__(self, constructor, *args, **keywords):
-        self.name = None
-        self.units = None
+    def __init__(self, constructor, *args, name=None, units=None, **keywords):
+        self.name = name
+        self.units = units
 
         self.frozen = None
         self.weight_decay = None

sknn/tests/test_native.py

@@ -6,15 +6,27 @@
 import numpy
 
 from sknn.mlp import Regressor as MLPR
-from sknn.mlp import Native as N
+from sknn.mlp import Native as N, Layer as L
 
 import lasagne.layers as ly
 import lasagne.nonlinearities as nl
 
 
 class TestNativeLasagneLayer(unittest.TestCase):
 
-    def test_DenseLinear(self):
-        nn = MLPR(layers=[N(ly.DenseLayer, num_units=4, nonlinearity=nl.linear)], n_iter=1)
+    def _run(self, nn):
         a_in, a_out = numpy.ones((8,16)), numpy.ones((8,4))
         nn.fit(a_in, a_out)
+
+    def test_DenseLinear(self):
+        nn = MLPR(layers=[N(ly.DenseLayer, num_units=4, nonlinearity=nl.linear)], n_iter=1)
+        self._run(nn)
+
+    def test_GaussianNoise(self):
+        nn = MLPR(layers=[L("Rectifier", units=12), N(ly.GaussianNoiseLayer), L("Linear")], n_iter=1)
+        self._run(nn)
+
+    def test_LeakyRectifier(self):
+        nn = MLPR(layers=[N(ly.DenseLayer, units=24, nonlinearity=nl.leaky_rectify),
+                          L("Linear")], n_iter=1)
+        self._run(nn)