Lasagne autoencoder: how do I just use the decoder part?

问题

Let's say I have an autoencoder in Lasagne, with two encoding layers, and two InverseLayers as a decoder:

input = InputLayer(...)
l1 = Conv1DLayer(input, ...)
l2 = DenseLayer(l1, ...)
# decoder part:
l2p = InverseLayer(l2, l2)
l1p = InverseLayer(l2p, l1)

and let's say I've trained this autoencoder to my satisfaction, and just wish to use the decoder; that is, I have data that I want to feed as an input to l2p (the first layer of the decoder part). How do I do this? I can't construct a network that consists just of the decoder part, it seems, because those are InverseLayers that depend on the existence of other layers.

来源：https://stackoverflow.com/questions/37147516/lasagne-autoencoder-how-do-i-just-use-the-decoder-part