Intel reportedly plans to release a software platform that simplifies and speeds up the training of AI models for computer vision.
The semiconductor giant has been said to be developing software for computer vision, among other things, in hopes of boosting chip sales by generating software revenue and becoming a bigger player in the chip market. AI.
(Photo: AFP/Getty Images)
New software that would simplify training AI models for computer vision
According to The Register, citing internal company documents, Intel plans to sell Sonoma Creek software as a subscription or license starting this fall as part of a revitalized push by the x86 giant to boost sales. of chips with a growing collection of commercial software products. Sonoma Creek is also seen as a way for Intel to increase the use of its OpenVINO open software toolkit for AI inference work.
Sonoma Creek, according to the chipmaker, will help various sectors. Computers can monitor assembly lines and look for manufacturing defects, and computers can discover abnormalities in medical imaging in healthcare. Smart agriculture, smart cities, retail and video security are some of the other businesses mentioned.
Sonoma Creek Operation
The registry said data collection, data labeling, model selection and training, model optimization and deployment are the five AI model training processes that would be scattered across multiple services in Sonoma Creek’s “unique workflow”.
Read more: The 5 most expensive apps in the App Store – More expensive than the iPhone?
Here’s how the software works:
First, Sonoma Creek will begin training computer vision models with as little as 10-20 photos or a video, using the PyTorch and TensorFlow frameworks.
Second, users could then use “intelligent assistants” to annotate the data, which helps Sonoma Creek learn and improve model accuracy.
Third, Sonoma Creek will automatically train the model and provide predictions once enough data has been annotated, which users can approve or modify to improve model accuracy.
Fourth, users can either use the model’s native framework for inference or use “Intel’s OpenVINO toolkit, which optimizes the model for a variety of Intel hardware, including CPUs and GPUs” .
Fifth, if a model is inaccurate, users can update it by adding new photos and identifying them before re-exporting the model. “Active learning” is the term used to refer to this process of constantly improving the model.
Sonoma Creek would be compatible with Intel and Nvidia processors
According to PCMag, Sonoma Creek would be compatible with Intel processors and Nvidia graphics cards (with plans to support Intel’s upcoming Arc graphics cards), as well as the company’s OpenVino toolkit and major software frameworks. machine learning PyTorch and TensorFlow.
Intel’s AI and software strategies will be bolstered by Sonoma Creek
Intel is developing Sonoma Creek in hopes of boosting chip sales by generating software revenue and becoming a bigger player in the AI market.
Intel is expanding its specialty chip portfolio, which will eventually include new discrete GPUs that the company hopes to ship for servers later this year while incorporating new AI features into its processors.
Related Article: Top 4 Best CPUs For 2022: Different Choices From AMD And Intel That Suit Your Needs