def random_play(n_steps): #env from example_3 height = 3 width = 3 start = [0, 0] goals = ([2, 2]) walls = None cliffs = None env = GridWorld(height, width, False, False, start, goals, walls, cliffs) #random actions over n_steps: env.reset() for step in range(n_steps): action = env.action_space_sample() new_state, reward, done = env.step(action) print("Step:", step, ", Action:", action, ", New state:", env.get_obs(), ", Done:", done, ", Reward:", reward) env.render(mode='episode')