Hello, I have a couple of doubts, after choosing the hyperparameters of the model by evaluating the cross-validation error, should the model be trained again with a new set formed by the training data plus the cross-validation data?
I also wonder if after having finished the whole process and even having evaluated the model with the test set at the end of all that, is it a good practice to retrain the model one last time using a new set formed by the set of training set, cross validation set and test set?