Displaying the data flow graph in TensorBoard

A practical way to obtain a visual display of your architecture and results is to have your own little TensorBoard_reader.py, as shown in the following lines of code:

def launchTensorBoard():
 import os
 os.system('tensorboard --logdir=' + './logs/tensorboard_logs')
 return

import threading
t = threading.Thread(target=launchTensorBoard, args=([]))
t.start()

#In your browser, enter http://localhost:6006 as the URL
#and then click on Graph tab. You will then see the full graph.

All you have to do is git the output directory to the log directory path in your program. It can be included in the main source code.

Once launched, you will see images of the previous section. You are now ready to enter corporate environments as a solid architect thinker and designer.

Use a TensorBoard reader such as this for your meetings. You will be viewed as an architect. You can use Microsoft PowerPoint or other tools naturally. But at one point, even a single view of this data flow graph, from a few seconds up to as long as your audience is interested, shows that you master the architecture of the solution. That is, you master the subject.