runningScore = 0 state = track.reset() runs = args.runs if args.train else 10 for run in range(runs): " for every lap or run (when training) run it. " score = 0 state = track.reset() maxTrackScore = 1000 for _track in range(maxTrackScore): " For every track out of the maximum score " if args.train: action, coefficient = actor.chooseActionTrain(state) else: action = actor.chooseAction(state) newState, reward, done, die = track.step(action * np.array([2., 1., 1.]) + np.array([-1., 0., 0.])) track.render() if args.train and actor.storeInBuffer( (state, action, coefficient, reward, newState)): print('updating model file') actor.update() score += reward state = newState