Cleanup

  >>> model = dc.models.Sequential(learning_rate=0.01)                  

  >>> model.add(layers.Dense(out_channels=2))                                  

  >>> model.add(layers.SoftMax())

  >>> model.fit(dataset, loss="binary_crossentropy", nb_epoch=1000)     

  >>> prediction = np.squeeze(model.predict_on_batch(X)) 

  """

  def __init__(self, **kwargs):