Intel boosts AI development with open source nGraph in Open Source connunis
Intel’s nGraph Compiler, a framework-neutral deep neural network (DNN) model compiler, is now open source. Data scientists working with TensorFlow on Intel Xeon Scalable processors can use Intel’s newly released simplified bridge code to deliver up to 10x performance improvements over previous TensorFlow integrations.
The goal is to make AI and deep learning development as accessible as possible. “Finding the right technology for AI solutions can be daunting for companies,” said – Arjun Bansal, vice president, AI Software, Intel. “With the nGraph Compiler, data scientists can create deep learning models without having to think about how that model needs to be adjusted across different frameworks, and its open source nature means getting access to the tools they need, quickly and easily,” he added.
The nGraph Compiler supports multiple deep learning frameworks while optimising models for multiple hardware solutions. The nGraph Compiler provides freedom of choice in frameworks and hardware to data scientists, says Intel. It lets framework owners add unique features with much less work, allows cloud service providers to more easily address a larger market demand, and helps enterprises maintain a consistent experience across frameworks and back ends, all without performance loss.
Currently, the nGraph Compiler supports three deep learning compute devices and six third-party deep learning frameworks. These are TensorFlow, MXNet, neon, PyTorch, CNTK and Caffe2. Intel will continue to add frameworks and devices in the coming months.
This news story is brought to you by softei.com, the specialist site dedicated to delivering information about what’s new in the electronics industry, with daily news updates, new products and industry news.
To stay up-to-date, register to receive our weekly newsletters and keep informed of the latest technology news and new products from around the globe. Simply click this link to register here: softei.com
Readers can also register to receive the monthly newsletter from our sister site, weartechdesign.com. This is the only website dedicated to developers of wearable electronic products. Simply click this link to register: weartechdesign.com