How to build a drag-and-drop-deep-learning model in 13 minutes (video)

Lobe just launched publicly last week.

Yes, this is complex stuff. But this video is fascinating no matter who watches…

Lobe is an easy-to-use visual tool that lets you build custom deep learning models, quickly train them, and ship them directly in your app without writing any code. Start by dragging in a folder of training examples from your desktop. Lobe automatically builds you a custom deep learning model and begins training. When you’re done, you can export a trained model and ship it directly in your app.

It’s a completely visual tool from designer Mike Matas and his co-founders Markus Beissinger and Adam Menges. I am always interested in anything Matas does, and Lobe is no exception.

You build and edit Lobe models through a web interface, and there’s a cloud API developers can use for finished models in production. But Lobe also exports to CoreML (for Apple platforms) and TensorFlow. My analogy: writing CoreML by hand is like writing PostScript by hand — possible, but only by a small number of talented experts. Lobe is to CoreML what Illustrator was to PostScript — a profoundly powerful tool that exposes the underlying technology to non-experts through an intuitive visual interface. Lobe looks utterly Matas-ian.

If you have any interest whatsoever in machine learning, drop what you’re doing right now and watch their 13-minute introductory tour. And if you’re not interested in machine learning, watch the video anyway and you’ll become interested in machine learning. It looks that amazing.

Via John Gruber at DaringFireball.net