def main_random(): env = AgarioEnv(render=RENDER, speed_scale=SPEED_SCALE, display_text=DISPLAY_TEXT, grid_resolution=GRID_RESOLUTION) agent = RandomAgent() for episode in range(NUM_EPISODES): state = env.reset() num_steps = 0 done = False while True: action = agent.get_action(state) for _ in range(NUM_SKIP_FRAMES): if RENDER: env.render() state, reward, done, _ = env.step(action) if done or num_steps >= MAX_STEPS: print(f'epoch: {episode}, max_mass = {state.mass}') agent.max_masses.append(state.mass) break num_steps += 1 agent.save_performance(path='random.performance') agent.print_final_stats() env.close()
def close(self): if (self.physics_connected): p.disconnect() self.physics_connected = False if __name__ == "__main__": env = BoosterBackEnv(render=True) epds = 3 #agent = LSTMAgent() #agent = MaxThrustAgent() agent = RandomAgent() #agent = DoNothingAgent() for epd in range(epds): obs = env.reset() done = False while not done: action = agent.get_action(obs) time.sleep(0.01) obs, reward, done, info = env.step(action) import pdb; pdb.set_trace()