Meta Teams with Qualcomm to Bring Llama 2 LLM to Mobile Devices
Meta will bring its new open-source Llama 2 generative AI model to mobile devices with the help of Qualcomm and its Snapdragon processors. The new open-source large language model will deploy on Qualcomm’s chips to enable on-edge processing, reducing reliance on cloud connectivity and improving the latency speed of the AI when accessed through smartphones and other consumer electronics beginning next year.
The Snapdragon on-device AI setup processes data locally instead of sending information to the cloud and relaying the results back to the user. That means developers can build more personalized experiences without sacrificing privacy, a major concern with many AI tools. Limiting the need for cloud communication also benefits developers and manufacturers by cutting costs and upping response times and overall performance. Qualcomm says it will eventually extend Llama 2 to PCs, virtual reality headsets, and even cars, maintaining the on-device approach for all of them.
Qualcomm Snapdragon platforms already handle a lot of AI data processing on devices, claiming more than 150 billion chips shipped. Ahead of Qualcomm releasing chips optimized for Llama 2, the company is encouraging developers to start optimizing their apps by using Qualcomm’s AI Stack software and hardware.
“We applaud Meta’s approach to open and responsible AI and are committed to driving innovation and reducing barriers-to-entry for developers of any size by bringing generative AI on-device,” Qualcomm senior vice president Durga Malladi said. “To effectively scale generative AI into the mainstream, AI will need to run on both the cloud and devices at the edge, such as smartphones, laptops, vehicles, and IoT devices.”