How to get started with Deep Learning for Vision Applications

EW ML main content (LC)

Data is transforming everything we do. And it will continue to re-shape technologies and businesses in the future, adding value in ways we couldn’t have previously imagined.

So it’s no surprise that organisations, from agile startups to OEMs and tech giants are focused on harnessing their data. But to succeed, they’ll need to add deep learning to their business analysis toolbox.

But if you’re new to this, where do you start?

There are plenty of off-the-shelf models available, and these are pretty neat, but they’re not always tailored to exactly what you need. Doing things from scratch is easier said than done and deploying it efficiently on an embedded platform can be even tougher.

Avnet Silica, in partnership with Deep Vision Consulting, has created a set of resources to help you create your own demo, and it’s easier than you think. All you’ll need is some Lego™ and our Deep Learning starter kit.

Step one: Watch the tutorial

Click here to watch full size video

To get started, watch this short step-by-step tutorial. And (this is very important), if you want to succeed, you’ll need to follow every single one of these steps, exactly as outlined in the video.

How to get started with Machine Learning - block diagram

Step two: Select your board

Order your preferred board by completing this form. We’ll send you the source code free of charge with your order. Already have a board? Tell us more about your project and we’ll send you the source files free of charge.

Xilinx Kria KV260 Vision AI Starter Kit - top side of the board

AMD-Xilinx Kria KV260 Vision
AI Starter Kit

Development platform for AMD/Xilinx Kria K26 SOM to built advanced vision applications without requiring complex hardware design knowledge
Access to Deep-Learning Processor Unit (DPU)
The kit includes a carrier board with an additional accessory kit for the power supply and cables

i.MX 8 Plus Development Board - angle view

NXP i.MX 8M Plus Evaluation Kit

Development kit for chip-down design on NXP i.MX 8M Plus
Chip includes a Neural Processing Unit (NPU)
The kit includes the base board, compute module and cables

Engicam EDIMM 2.0 - top side of the board

Engicam i.Core MX8M Plus EDIMM 2.0 Starter Kit

System-on-Module (SOM) with NXP i.MX 8M Plus with a Neural Processing Unit (NPU)
Optimized Board Support Package
The kit includes a carrier board with industrial connectivity options, a 7-inch touch display, and cables