Closed isaienkov closed 4 years ago
Is there any possibility to save superlearner with a KerasRegressor inside it? When I use ensemble without KerasRegressor it works fine like:
import pickle pickle.dump(model, open(filename, 'wb'))
But if we have KerasRegressor in any layer it is not working.
Hi! This is a problem with TensorFlow not being serializable. It might help if you switch the Keras backend (#97)?
Is there any possibility to save superlearner with a KerasRegressor inside it? When I use ensemble without KerasRegressor it works fine like:
But if we have KerasRegressor in any layer it is not working.