以下代码尽管无法运行，但可供阅读参考。

class 
    def __init__(self):

        self.sess = tf.Session()

        self.summary()

        self.sess.run(tf.global_variables_initializer())   

    def summary(self):
        '''
        First, define summary_merged and summary_write.
        '''

        self.summary_merged = tf.summary.merge_all()

        self.summary_write = tf.summary.FileWriter(self.work_path + 'log/BC/%s' % time.strftime('%m%d%H%M%S', time.localtime()), graph=self.sess.graph)

    def build_graph(self):

        tf.summary.scalar('bc_loss', tf.reshape(self.bc_loss, []))
        '''
        Add any scalar you want to monitor.
        '''

    def train_step(self, state, action):

        fetches = {
            'summary': self.summary_merged,
        }
        '''
        Second, run the summary_merged with sess.
        '''

    def train(self):
        train_step = 0
        for _ in range(50):
            train_step += 1

            i_train = train_step
            self.summary_write.add_summary(results['summary'], i_train)
            '''
            Third,
            Use summary_write here to add summary.
            Do not forget set i_train.
            '''

TO DO