MediaTek, a semiconductor company based in Taiwan, is leveraging Meta’s generative AI platform for edge devices. The company aims to establish a comprehensive edge computing ecosystem that fast-tracks the development of artificial intelligence (AI) applications using Meta’s Llama 2, an open-source Large Language Model (LLM).
Utilizing Meta’s generative AI platform, MediaTek will employ its APUs and NeuroPilot AI Platform to enhance various edge devices such as smartphones, Internet of Things (IoT) devices, smart vehicles, smart homes, and more.
“The increasing popularity of Generative AI is a significant trend in digital transformation, and our vision is to provide the exciting community of Llama 2 developers and users with the tools needed to fully innovate in the AI space,” said JC Hsu, corporate senior VP and GM of Wireless Communications Business Unit at MediaTek.
One of the advantages of utilizing Llama 2 models is that it can execute AI applications directly on the device, rather than relying on cloud-based processing. This approach offers developers seamless performance, heightened privacy, improved security and reliability, reduced latency, operation in low-connectivity areas, and decreased operational costs.
More edge capabilities
MediaTek said that all its shipped 5G smartphone SoCs come equipped with APUs designed to handle a wide array of generative AI features, including AI Noise Reduction, AI Super Resolution, AI MEMC, and more.
The company’s upcoming flagship chipset, set to be launched later this year, will feature a software stack optimized for Llama 2 execution. The upgraded APU in this chipset will incorporate Transformer backbone acceleration, reduced footprint access, and improved utilization of DRAM bandwidth. These enhancements will further elevate the performance of LLM and AIGC (AI-generated content).
“Through our partnership with Meta, we can deliver hardware and software with far more capability on the edge than ever before,” Hsu said.
MediaTek expects Llama 2-based AI applications to become available for smartphones powered by the next-generation flagship SoC, scheduled to hit the market by the end of the year.