Exclusive Intel plans to start selling a software platform that promises to simplify and speed up the training of AI models for computer vision, according to internal company documents seen by The register.
The software, named Sonoma Creek, is set to launch this fall, and the semiconductor giant plans to sell it as a subscription or license, part of a new push by the x86 giant to boost sales of chips with a growing stable of commercial software products. Intel is also looking at Sonoma Creek as a way to drive adoption of its free OpenVINO software toolkit for AI inference work.
Intel touts Sonoma Creek as an “end-to-end AI development platform” that simplifies computer vision model training for subject matter experts without a data science background.
The chipmaker believes Sonoma Creek will benefit multiple industries. This includes manufacturing, where computers can monitor assembly lines and find defects, and healthcare, where computers can detect abnormalities in medical imaging. Other industries cited include smart agriculture, smart cities, retail and video security.
The Sonoma Creek development has been hidden for some time. However, there have been some public references to the software. This includes a publicly available online workshop that Intel hosted for users in South Korea last year, in which a company representative discussed the software. The CEO of Intel’s Korean operations also mentioned Sonoma Creek in a Korean press article in December.
How Sonoma Creek Works
In a presentation, Intel says Sonoma Creek brings together five important stages of AI model training that would be fragmented across multiple services into a “single workflow.” These five steps are data collection, data labeling, model selection and training, model optimization, and deployment.
A slide from a workshop in South Korea showing Intel’s perspective on fragmented AI development. Click to enlarge.
The software works like this, according to several documents:
A slide from a workshop in South Korea showing how Intel’s Sonoma Creek software addresses the five stages of AI model development. Click to enlarge.
Sonoma Creek begins training computer vision models using PyTorch and TensorFlow frameworks with just 10-20 images or a video.
Users then annotate the data using what Intel calls “intelligent assistants,” which helps Sonoma Creek learn and improve the accuracy of the model.
Once enough data is annotated, Sonoma Creek will automatically train the model and make predictions, which users approve or modify to improve model accuracy as well.
After that, users export the model for inference using the original model framework or using Intel’s OpenVINO toolkit, which optimizes the model for a variety of Intel hardware, including CPUs and GPUs.
If a template is not accurate, users can make corrections by adding new images and tagging those images before re-exporting the template. This process of continuous improvement of the model is called “active learning”.
In an internal explainer for Intel employees, the company says Sonoma Creek comes with an intuitive web-based graphical user interface. This, combined with the annotation helpers, “helps[s] speed up data annotation and reduce[s] deployment time,” according to Intel. Sonoma Creek also allows users to “combine multiple tasks,” such as detection and classification, “for more granular model development.”
Intel says these and other features “differentiate the platform from other” no-code or low-code development platforms that have become popular.
Sonoma Creek runs on Intel processors and Nvidia GPUs for desktops and servers, although Intel plans to support its own GPUs in the future.
A slide from a workshop in South Korea illustrating how Intel’s Sonoma Creek software works. Click to enlarge.
Intel began prototyping Sonoma Creek with early adopters in the second half of last year, and the software has been in an early access program for a few months, according to a document. This is all part of the company’s plan to launch the software as a commercial offering this fall.
The x86 giant plans to sell Sonoma Creek as a subscription service or software license. The software will be available for on-premises servers and through cloud service providers. For cloud deployments, users will be able to access the software in virtual machines or Kubernetes containers.
Intel also plans to provide support and professional services for the suite, which includes “onsite solution activation, project-level consultation, and enterprise-level support.”
One of the go-to-market channels Intel is considering is system manufacturers, which is how the company has already sold some of its current software products.
Sonoma Creek will boost Intel’s AI and software strategies
Intel is working on Sonoma Creek as the company hopes to boost chip sales through software revenue and become a bigger player in the AI space.
The x86 giant is trying to catch up with Nvidia, which has built a comprehensive portfolio of AI hardware and software over the past few years, making it a favorite on Wall Street.
On the hardware front, Intel is taking several avenues, integrating new AI features into its processors, while expanding its portfolio of specialty chips, which will eventually include new discrete GPUs the company plans to release for servers later this year. year.
The chipmaker has also started selling AI software products in addition to free offerings like OpenVINO. These commercial products include the managed AI service Cnvrg.io and the AI model optimization platform SigOpt.
As one document states, Intel sees Sonoma Creek as the “next level” in the company’s AI strategy. The big questions are whether enough organizations will find the software useful and whether it will live up to the “ecosystem-friendly” promise that Intel CEO Pat Gelsinger has made.
Sonoma Creek will join a growing stable of commercial software offerings that Gelsinger says will help Intel better compete with rivals by making its hardware more attractive. After earning $100 million on software last year, the business hopes to increase that figure to $150 million this year.
That $150 million figure may pale in comparison to the $76 billion Intel expects to collect in sales this year, though Gelsinger promised software would have an outsized impact on the business.
“We’re going to grow it quickly, but we’re going to get platform differentiation right, which is going to [have a] very disproportionate business advantage, even if the short-term earnings are rather low,” he told an investor meeting in February. ®