fix sub_pixel_upscaling test

tests/keras_contrib/layers/test_convolutional.py

@@ -156,9 +156,11 @@ def test_sub_pixel_upscaling():
     num_samples = 2
     num_row = 16
     num_col = 16
+    input_dtype = K.floatx()
 
     for scale_factor in [2, 3, 4]:
         input_data = np.random.random((num_samples, 4 * (scale_factor ** 2), num_row, num_col))
+        input_data = input_data.astype(input_dtype)
 
         if K.image_data_format() == 'channels_last':
             input_data = input_data.transpose((0, 2, 3, 1))