Fails to load TensorFlowModelDataset with SSIM metric

[eileen-kuehn]

The current implementation in #1 fails when training with tensorflow as the saved model cannot be correctly loaded since the SSIM metric has been introduced:

DataSetError: Failed while loading data from data set 
TFPTModelDataset(filepath=/.../qc_data_preparation/data/06_models/autoencoder_1-2-3-4, protocol={}, 
save_args={'save_format': tf}).
Unable to restore custom object of class "MeanMetricWrapper" (type _tf_keras_metric). Please make sure that this class is included in the 
`custom_objects` arg when calling `load_model()`. Also, check that the class implements `get_config` and `from_config`.

Complete metadata: {'class_name': 'MeanMetricWrapper', 'name': 'ssim', 'dtype': 'float32', 'config': {'name': 'ssim', 'dtype': 'float32'}, 
'shared_object_id': 2}