5 Open Source AI platforms to try
The sound of Open Source AI music
I would say that I have become more and more taken by Open Source (and all that it brings) in recent time, and I definitely intend to explore its myriad opportunities more intensely. I have also adjusted to the inevitability of Artificial Intelligence, or A.I. - that branch of computer science that develops algorithms (step-by-step instructions to solve a problem) and programs that help give machines the illusion of operating in more human-like ways - and how it can actually serve an impressive amount of good.
But the A.I. that's had all the major reviews so far have not been Open Source. But, that
shouldn't be. Right?
One reason may be the enormous resources required to set things up.
But, shouldn't that be where Open Source can actually help?
Well, I found a handful of A.I. projects
that should satisfy your itch to combine those 2.
5 Open Source AI platforms to try
The following tools/platforms may be just what you'll need if you're starting out, wanting to learn how it all works or cobbling out a project to test a business idea. These are listed in no particular order, but (the last time I checked) are free to start developing with:
1. OpenCV
This may actually deserve the top spot, if I were ranking, for what it offers. It describes itself as "an open source computer vision and machine learning software library", and reportedly has been used by well-established companies and startups to do things that span from intrusion detection in surveillance videos in Israel, detecting swimming pool accidents in Europe, putting streetview images together, and checking runways for debris in Turkey.
With an Apache 2 license, it claims to have more than 2500 optimized algorithms, and make it easy for businesses to use and modify the code.
Favourite thing: Computer Vision applications development, supports Windows, Linux, Android, and Mac OS, with C++, Python, Java, and MATLAB interfaces.
2. H2O.ai
This company believes that A.I. should be democratized. That there should be freedom around the creation and use of A.I. It's said to have begun as a "grassroots effort" and is a community of open source contributors, business leaders, nonprofits, and academics.
It prides itself as being "a global visionary and thought leader in..responsible AI".
Favourite thing: AI cloud platform, appears to be aware of ethical issues in AI.
3. Tensorflow
Self-described as an end-to-end platform that makes it easy to build and deploy Machine Learning models.
It provides tutorials, examples, and other resources to speed up model building and create scalable ML solutions.
Favourite thing: Community, deploy to web (tensorflow.js), mobile (tensorflow lite), and servers (tfx), learning tutorials.
4. Keras
Keras is said to be an "API designed for human beings, not machines".
Paging Sarah Conor
And it wants to give an "unfair advantage" to any developer looking to ship Machine Learning-powered apps. Something that should resonate with indie-devs.
Built on top of the Tensorflow platform, Keras enables you to export models to run directly in the browser (with JavaScript), on iOS, Android, or embedded devices (with TF Lite), even via a web API.
Favourite thing: Vast ecosystem, deploy anywhere, state of the art research (they say it's used by CERN and NASA).
5. PyTorch
PyTorch is similar to Tensorflow in being an entire framework that supports various aspects of open source AI development, with vast libraries and datasets.
Easy to use for developers who already code with the Python programming language, but it also has a C++ interface.
Favourite thing: (For those who love Python) uses Python, tutorials.
The revolution will be "Edelweiss" 🎤🎻
Artificial Intelligence can help new (and, of course, not so new) developers create products of great value. And not facing the restrictions which closed source quarters bring is most desirable. If you are a beginner dev looking for your first paid position, but faced with a bare portfolio or zero production experience, using Open Source AI may very well help to impress at your next job application.
Explore these 5 tools, and let your imagination come alive with something, that brings value, to build.
But remember, for all the power that AI brings, it also brings a pachydermous amount of responsibility.
Good luck.