Training a model in new example set

philo

Hello,
can anyone tell me if  it is possible to train a  model in a new example set.
I know that this is trivial if you know before hand the type of the model.
However, I am using a X-Validation to select between different types of models with different configurations of their parameters
and thus I do not know what will the outcome be.
So, I have a model as an output and I am struggling to train it again.
In general , I would suggest to include an operator of this type in the future releases ,since I find it very useful.
Thanks in advance.