tf.logging用法

用来记录模型训练过程的一些参数。

常用的几项：

tf.logging.set_verbosity(tf.logging.INFO)
设计日志级别. 控制那些日志打印到屏幕上。

tf.logging.info(msg, *args, **kwargs)
记录INFO级别的日志. args 是配合msg中的占位符用的. 比如 info("I have been in love with %s for %d years.","yichu",7)

tf.logging.log_every_n( level, msg, n, *args)
改行代码每执行n次输出一次. 打印的时机分别是(1, n+1, 2n+1,…).
 

举例说明：

def main():
    tf.logging.set_verbosity(tf.logging.INFO)
    tf.logging.info("I have been in love with %s for %d years.", "yichu", 7)
    # 因为set_verbosity这里只设置了INFO级别的日志，所以只会打印上一条I have...years这一条，
    # 不会打印下面这一行this is ... info这一条，如果希望输出INFO和DEBUG2个级别的信息，
    # 那么需要再次调用tf.logging.set_verbosity(tf.logging.DEBUG)才可以打印info和debug2个级别。
    tf.logging.debug("this is a debug info")

    train_op = tf.train.GradientDescentOptimizer(learning_rate=LEARNING_RATE).\
        minimize(loss_tensor, global_step=tf.train.create_global_step())
    with tf.Session() as sess:
        ...
        tf.logging.log_every_n(tf.logging.INFO, "np.mean(loss_evl)= %f at step %d", 100, np.mean(loss_evl),
                               sess.run(tf.train.get_global_step()))

    # 注意，最后的tf.logging.log_every_n是每隔100次打印一次，打印的日志如下：
    # INFO:tensorflow:np.mean(loss_evl)= 1.396970 at step 1
    # INFO:tensorflow:np.mean(loss_evl)= 1.221397 at step 101
    # INFO:tensorflow:np.mean(loss_evl)= 1.061688 at step 201

 

 

参考博客：

https://blog.csdn.net/chuchus/article/details/79726446