Intel teams up with Accenture to launch 34 open-source AI reference kits

In a powerful collaboration with Accenture, Intel has rolled out an impressive collection of 34 open-source AI reference kits. These kits are potential game-changers, aimed at simplifying and speeding up the process of deploying AI for data scientists and developers.

Think about it, dear readers -- every kit is a treasure chest of AI tools. Inside, you'll find model code, training data, instructions for setting up the machine learning pipeline, libraries, and oneAPI components. All of these are designed to optimize AI and make it more accessible, regardless of whether you're working in an on-premises, cloud, or edge environment.

The magic behind these kits is Intel’s end-to-end AI software portfolio and the open, standards-based oneAPI programming model. What's the result? Developers can introduce AI into their applications more smoothly, bolster their intelligent solutions, and deploy everything faster. It's like taking the express lane instead of the traditional model development workflow.

What's exciting is that these kits aren't one-size-fits-all. They've been preconfigured to meet the specific needs of a range of industries. So whether you're in consumer products, energy, utilities, financial services, health and life sciences, manufacturing, retail, telecommunications, or something else, there's a kit for you.

Let's talk about some of the benefits we've seen so far. One kit is designed for enterprise conversational AI chatbot interactions. And guess what? Users have seen inferencing speeds in batch mode that are up to 45 percent faster, thanks to oneAPI optimizations.

Then there's a kit that's been made for automating visual quality control inspections in the life sciences sector. The results have been impressive, with training up to 20% faster and inferencing 55 percent faster for visual defect detection.

For those in utilities, there's a kit that has shown a 25 percent increase in prediction accuracy for utility asset health. Now that's something that can really make a difference in delivering reliable service.

What's amazing about these AI reference kits is the potential time reduction, going from weeks to just a few days. Developers and data scientists can train models faster and more affordably, without the constraints of proprietary environments. Plus, with the portability of AI tools and optimizations powered by oneAPI, we're looking at a new era of open accelerated computing applications.

John Giubileo, Accenture’s managing director, couldn't be happier. "Working with Intel on these AI reference kits has boosted productivity for our clients' AI workloads," he said. "Thanks to oneAPI, these kits offer a portable, efficient solution for AI projects that cuts down on complexity and speeds up deployment across all industries."

Looking to the future, Intel plans to continue updating select kits based on community feedback and contributions. Among these kits, you'll find ones for visual quality inspection, setting up enterprise conversational AI chatbots, predictive asset health analytics, medical imaging diagnostics, document automation, and AI-structured data generation. The best part? You can download any of them for free, either on the Intel web page or GitHub.

Image credit: lightsource/depositphotos.com