By Dante Sblendorio
TensorFlow is an open source machine learning library used by many companies within the industry to perform numerical computation. It was originally developed by engineers working within Google’s AI team. It has an extremely flexible architecture and can be deployed across multiple types of platforms, from mobile phones to clusters of servers. It can run on CPUs, GPUs, and even specifically designed machine learning hardware (ASICs). Both Python and C APIs are released, in addition to Java and Go. There are also third-party packages available for many other languages.
In this article, we’ll get started with TensorFlow by walking through the process for installing it. You can use different languages with TensorFlow, but we’ll be using Python, which is a good language to start with if you’re new to programming.
There are several options for installing TensorFlow. The best option depends on your operating system, hardware, and environment preferences. I would recommend installation via Anaconda, as it works across all platforms (Linux, Mac OS X, Windows), and is easy and simple to use. It is an open source package management system and environment management system for data analysis and scientific computing. Installation instructions for Anaconda can be found here. We will be using Python, although Anaconda also supports package and environment management for R, Node.js, Java and other application stacks. After installing Anaconda, TensorFlow can be installed by following these steps:
To create a conda environment tensorflow, open the Anaconda prompt and run:
conda create -n tensorflow pip python=3.5
Enter the environment by running:
conda activate tensorflow
Once inside your environment, call the command to install TensorFlow. Run:
pip install --ignore-installed --upgrade tensorflow
I’ll only include installation for CPU versions of TensorFlow. For those new to TensorFlow, this is more than sufficient. As you become more familiar with the package, and speed and processing power become more critical, the GPU version may be more formidable, permitting you have the supporting hardware.
The contest entry code for challenge 1 is your CodeProject member number. You can find this on your CodeProject profile page. Please click here to enter the contest entry code.