TF Learning – Tensorboard for TF: Tensorboard visualization introduction, introduction, and detailed walkthrough of how to use Tensorflow

TF Learning – Tensorboard for TF: Tensorboard visualization introduction, introduction, and detailed walkthrough of how to use Tensorflow

 

directory

Tensorboard profile

Tensorboard introduction to various sections

Tensorboard usage

---

 

 

 

Tensorboard profile

TensorBoard is a very cool feature in TensorFlow. As we all know, neural network is often like a black box. What is inside, what is the structure, how is the training, can be difficult to understand. The purpose of TensorBoard is to visualize the complex neural network training process, and to better understand, debug, and optimize the program.

 

 

Tensorboard introduction to various sections

1.

2,

3,

4,

 

 

Tensorboard usage

First, generate the logs folder and its files

$tensorboard --logdir= logs $tensorboard --logdir= logs
merged = tf.summary.merge_all()
writer = tf.summary.FileWriter("logs", sess.graph)
sess.run(tf.global_variables_initializer())
Copy the code

 

2. Open the webpage

Tensorboard –logdir=logs http://laptop-tvio85p6:6006/#scalars or HTTP :/127.0.0.1:6006 or http://localhost:6006

 

 

 

Related articles

TF: Tensorboard practice: Get the neural network Tensorboard form to events.out. tfEvents file + run the file in DOS local server output to webpage visualization TF: SCALARS, GRAPHS, constant, HISTOGRAMS of Tensorboard