Reconstruct or inverse transform from Feature in UnsupervisedDBN

[mgcy]

I have fitted and transformed an model as:
X_feature = unsupervised_dbn.fit_transform(X)

How can I inverse transform the X_feature to get the reconstructed X_re?

[larasatiayudia]

i dont know if its right but logically you can do this:
feature=dbn.transform(data)
reconstructed=copy.copy(feature)
for i in range(len(dbn.rbm_layers),0,-1):
reconstructed = dbn.rbm_layers[i-1]._reconstruct(reconstructed)