Fix keras frontend elementwise-ops for lists with len>2 (fixes issue #325) (#326)

* Added elementwise-add test

* Fix typo

* Fixed elem-wise ops for lists with len>2

nnvm/python/nnvm/frontend/keras.py

@@ -77,18 +77,21 @@ def _convert_advanced_activation(insym, keras_layer, _):
 
 def _convert_merge(insym, keras_layer, _):
     merge_type = type(keras_layer).__name__
+    ret = insym[0]
+    for i in range(1, len(insym)):
         if merge_type == 'Add':
-        return _sym.elemwise_add(insym[0], insym[1])
+            ret = _sym.elemwise_add(ret, insym[i])
         elif merge_type == 'Subtract':
-        return _sym.elemwise_sub(insym[0], insym[1])
+            ret = _sym.elemwise_sub(ret, insym[i])
         elif merge_type == 'Multiply':
-        return _sym.elemwise_mul(insym[0], insym[1])
+            ret = _sym.elemwise_mul(ret, insym[i])
         elif merge_type == 'Average':
             raise NotImplementedError('Average merge not implemented')
         elif merge_type == 'Maximum':
             raise NotImplementedError('Maximum merge not implemented')
         else:
             raise TypeError("Unsupported merge type : {}".format(merge_type))
+    return ret
 
 
 def _convert_dense(insym, keras_layer, symtab):

nnvm/tests/python/frontend/keras/test_forward.py

@@ -39,6 +39,27 @@ def verify_keras_frontend(keras_model):
         np.testing.assert_allclose(keras_out, tvm_out, rtol=1e-5, atol=1e-5)
 
 
+def test_forward_elemwise_add():
+    print("test_forward_elemwise_add")
+    r = []
+    data = keras.layers.Input(shape=(32,32,3))
+    x = keras.layers.Conv2D(8, (3, 3), padding="same")(data)
+    r.append(x)
+    x = keras.layers.Conv2D(8, (3, 3), padding="same")(x)
+    r.append(x)
+    x = keras.layers.Conv2D(8, (3, 3), padding="same")(x)
+
+    y = keras.layers.add([keras.layers.add([x, r[0]]), r[1]])
+    y = keras.layers.GlobalAveragePooling2D()(y)
+    keras_model = keras.models.Model(data, y)
+    verify_keras_frontend(keras_model)
+
+    y = keras.layers.add([x, r[0], r[1]])
+    y = keras.layers.GlobalAveragePooling2D()(y)
+    keras_model = keras.models.Model(data, y)
+    verify_keras_frontend(keras_model)
+
+
 def test_forward_softrelu():
     data = keras.layers.Input(shape=(32,32,3))
     x = keras.layers.Activation('softplus')(data)
@@ -114,6 +135,7 @@ def test_forward_resnet50():
 
 
 if __name__ == '__main__':
+    test_forward_elemwise_add()
     test_forward_softrelu()
     test_forward_leaky_relu()
     test_forward_dense()