How to use the openai.deepq.build_train function in openai

updates_queue: multiprocessing.Queue,
        weights_queue: multiprocessing.Queue,
        network, seed, config, lr, total_timesteps, learning_starts,
        buffer_size, exploration_fraction, exploration_initial_eps, exploration_final_eps,
        train_freq, batch_size, print_freq, checkpoint_freq, gamma,
        target_network_update_freq, prioritized_replay, prioritized_replay_alpha,
        prioritized_replay_beta0, prioritized_replay_beta_iters,
        prioritized_replay_eps, experiment_name, load_path, network_kwargs):
    _ = get_session()
    set_global_seeds(seed)
    q_func = build_q_func(network, **network_kwargs)

    def make_obs_ph(name):
        return ObservationInput(DotaEnvironment.get_observation_space(), name=name)

    _, train, update_target, debug = deepq.build_train(
        scope='deepq_train',
        make_obs_ph=make_obs_ph,
        q_func=q_func,
        num_actions=DotaEnvironment.get_action_space().n,
        optimizer=tf.train.AdamOptimizer(learning_rate=lr),
        gamma=gamma,
        grad_norm_clipping=10,)

    if prioritized_replay:
        replay_buffer = PrioritizedReplayBuffer(buffer_size, alpha=prioritized_replay_alpha)
        if prioritized_replay_beta_iters is None:
            prioritized_replay_beta_iters = total_timesteps
        beta_schedule = LinearSchedule(prioritized_replay_beta_iters,
                                       initial_p=prioritized_replay_beta0,
                                       final_p=1.0)
    else: