Qualcomm Technologies has unveiled its Qualcomm AI Stack portfolio, bringing a unified software approach to creating AI applications across multiple different Qualcomm products, a move that comes as many semiconductor technology companies are going “full stack” to help developers accelerate what otherwise could be complex and clunky creative activities.
The AI Stack portfolio combines and improves Qualcomm’s existing AI offerings, the company said, by delivering a comprehensive AI solution for OEMs and developers, supporting a wide range of intelligent devices with broad AI software access and compatibility. It enables a single AI software portfolio that works across Qualcomm-powered Connected Intelligent Edge products, including mobile, automotive, extended reality, computing, IoT, and cloud platforms.
Qualcomm CEO Cristiano Amon has made references in the past to how he wants to transform Qualcomm into a “full stack” company addressing opportunities at the intelligent edge, and this certainly seems to be a step in that direction.
The new portfolio provides direct access to the Qualcomm AI Engine and dedicated AI cores on Qualcomm Cloud AI 100 using Qualcomm AI Engine direct, the company said. As part of the new offering, the Qualcomm AI Engine direct will now scale across every AI accelerator inside a comprehensive range of Qualcomm Technologies products. Qualcomm AI Engine direct is an AI library that delegates and deploys existing models directly to the AI accelerators on Qualcomm Technologies platforms and will now provide OEMs and developers the ability to develop one feature, then move the same model across different products and tiers.
The Qualcomm AI Stack also supports different AI frameworks and popular runtimes, including TensorFlow, PyTorch and ONNX, developer libraries and services, system software, tools, and compilers so that any AI feature developed for one device can easily be deployed on others.
Yuan Yu, Distinguished Engineer for AI Frameworks Architecture at Microsoft, applauded the move, saying in a statement, “Qualcomm Technologies continues to deliver advanced AI solutions that make it easier for developers to leverage today’s most powerful AI and ML capabilities. We’re enthusiastic to bring Qualcomm AI Software technologies to developers in our ecosystems, including the Qualcomm Neural Processing SDK for Windows, to offer edge-to-cloud AI capabilities by providing unified, comprehensive AI software that scales from clients to servers.”
The Qualcomm AI Stack portfolio will also include a suite of tools, including Qualcomm AI Model Efficiency Toolkit (AIMET), AI development graphical user interface (GUI), model analyzers to foster enhanced quantization and optimization, and Neural Architecture Search (NAS).