def do_trial(): global observation, reward, t, img, action, done, info, avg_reward observation = env.reset() net.Flush() f = 0 for t in range(300): if render_during_training: time.sleep(0.01) env.render() # interact with NN inp = interact_with_nn() if render_during_training: img = viz.Draw(net) cv2.imshow("current best", img) cv2.waitKey(1) action = np.argmax(out) observation, reward, done, info = env.step(action) if done: break f += reward avg_reward += f return avg_reward
def do_trial(env, net, render_during_training): observation = env.reset() net.Flush() f = 0 for t in range(500): if render_during_training: #time.sleep(0.001) env.render() # interact with NN interact_with_nn(env, net, t, observation) if render_during_training: img = viz.Draw(net) cv2.imshow("current best", img) cv2.waitKey(1) action = np.array(out) observation, reward, done, info = env.step(action) f += reward return f
try: observation = env.reset() net = NEAT.NeuralNetwork() g = pickle.loads(hof[-1]) g.BuildPhenotype(net) reward = 0 for t in range(250): time.sleep(0.01) env.render() # interact with NN interact_with_nn() # render NN img = viz.Draw(net) cv2.imshow("current best", img) cv2.waitKey(1) action = np.argmax(out) observation, reward, done, info = env.step(action) if done: break except Exception as ex: print(ex) time.sleep(0.2)