How to use the stable-baselines.stable_baselines.a2c.a2c.A2C function in stable-baselines

self.actions_ph = None
        self.advs_ph = None
        self.rewards_ph = None
        self.pg_loss = None
        self.vf_loss = None
        self.entropy = None
        self.apply_backprop = None
        self.train_model = None
        self.step_model = None
        self.proba_step = None
        self.value = None
        self.initial_state = None
        self.learning_rate_schedule = None
        self.summary = None

        super(A2C, self).__init__(policy=policy, env=env, verbose=verbose, requires_vec_env=True,
                                  _init_setup_model=_init_setup_model, policy_kwargs=policy_kwargs,
                                  seed=seed, n_cpu_tf_sess=n_cpu_tf_sess)

        # if we are loading, it is possible the environment is not known, however the obs and action space are known
        if _init_setup_model:
            self.setup_model()