Tensorboard 可视化

1. 将想要命名的值添加代码及缩进

with tf.name_scope('input'):
    with tf.name_scope('train_data'):
        ph_train_data = tf.placeholder(tf.float32, [None, n_time_step, n_dim])
    with tf.name_scope('train_label'):
        ph_train_label = tf.placeholder(tf.float32, [None, n_classes])

2. 运行 session

init = tf.global_variables_initializer()
with tf.Session() as sess:
    sess.run(init)
    tf.summary.FileWriter("logs/", sess.graph)

3. 运行代码后，graph数据将自动保存在当前项目下的 "logs" 文件夹下

4. 打开cmd，文件夹转到要打开的 logs 所在项目，输入 tensorboard --logdir logs

复制网址到Google浏览器打开 （或者直接http://localhost:6006）即可看见生成的 Graph