Hi Gregory,
cross-validation is just for evaluation purposes; it's not meant to create a model.
Usually, you would build your model with all the training data available. Then, however, you don't have unseen test data to evaluate your model. This is where cross-validation comes into play -- it's purpose is to evaluate your model, not to build one.
Best
Max