Correctly load state from pretraining

[davnn]

    # Filter out decoder network keys
    ae_net_dict = {k: v for k, v in ae_net_dict.items() if k in net_dict}

Did not match any keys. The encoders seem to live in .encoder of the corresponding AE and we can directly import the keys from there.

[davnn]

addresses https://github.com/lukasruff/Deep-SAD-PyTorch/issues/9