TensorFlow Getting Started

TensorFlow is an end-to-end open source machine learning platform. It has a comprehensive and flexible ecosystem, which contains a variety of tools, libraries and community resources. In terms of machine learning, it can easily build models, carry out reliable machine learning production anytime and anywhere, and conduct powerful research experiments.

Preparation

1. Get the Internet IP of your Server on Cloud
2. Check your Inbound of Security Group Rule of Cloud Console to ensure the TCP:6006,80 is allowed
3. Complete Five steps for Domain if you want to use Domain for TensorFlow
4. Get default username and password of TensorFlow

TensorFlow Initialization

Steps for you

1. Use the browser of your local computer to access the URL: http:// domain name or http:// server public IP to enter the login page
   

2. Use SSH to connect Server, run the following command to obtain the token:

   $ docker exec -it tensorflow jupyter notebook list
   Currently running servers:
   http://0.0.0.0:8888/?token=c8929544462391e32bbf0d7763b7b5dda3ab00b2f14da5b9 :: /tf
   

3. Log in to the console and use jupyter (edit the source code)
   

4. Use SSH to connect Server, run the following command to start TensorBoard

   $ docker exec -it tensorflow bash
   $ cd /usr/local/lib/python3.8/dist-packages/tensorboard && tensorboard --logdir=/data/logs --port 6006 --host 0.0.0.0

5. Using local Chrome or Firefox to visit the URL http://DNS:6006 or http://Internet IP:6006 to access TensorBoard

More useful TensorFlow guide, please refer to TensorFlow Documentation

Having trouble?

Below is for you to solve problem, and you can contact Websoft9 Support or refer to Troubleshoot + FAQ to get more.

TensorFlow QuickStart

Now, we run a TensorFlow sample for quick start:

TensorFlow Setup

TensorBoard Password

Reference sheet

The below items and General parameter sheet is maybe useful for you manage TensorFlow

Run docker ps command, view all Containers when TensorFlow is running:

CONTAINER ID   IMAGE                                  COMMAND                  CREATED        STATUS         PORTS                                                                                  NAMES
f7e151917bec   tensorflow/tensorflow:latest-jupyter   "/bin/bash -c 'cd /u…"   15 hours ago   Up 2 minutes   0.0.0.0:6006->6006/tcp, :::6006->6006/tcp, 0.0.0.0:9001->8888/tcp, :::9001->8888/tcp   tensorflow

Path

TensorFlow installation directory： /data/apps/tensorflow
TensorFlow notebooks directory： /data/apps/tensorflow/data/tensorflow

Port

Port	Use	Necessity
6006	Access TensorBoard via HTTP	Optional

Version

docker exec -it tensorflow grep "_VERSION =" /usr/local/lib/python3.8/dist-packages/tensorflow/tools/pip_package/setup.py| cut -d= -f2

Service

sudo docker start | stop | restart tensorflow

CLI

TensorFlow provides a powerful command-line tool 'tfx', which can be installed by executing the following commands:

source /data/apps/tensorflow/bin/activate
pip install tfx

tfx is only supported until 2.3.2, and installation may cause the TensorFlow version to be downgraded

API

API Documentation