The module that runs a training (runs the episodes and updates the model) is hard-wired to out q-learning tabular method. We should program to a model abstraction here, so we can reuse the trainer code to train with the tabular model as well as the neural net
The module that runs a training (runs the episodes and updates the model) is hard-wired to out q-learning tabular method. We should program to a model abstraction here, so we can reuse the trainer code to train with the tabular model as well as the neural net