首页 News 正文

During the Spring Festival of the the Year of the Loong in China, OpenAI released the video generation model Sora, known as the "World Simulator", which has attracted extensive attention and discussion. In recent years, various AI models have emerged, from text dialogue robots to text generated images, text generated videos, etc., and generative AI has shown a strong development trend. Research data released by McKinsey shows that generative AI technology can achieve an overall economic growth of $2.6 trillion to $4.4 trillion per year for over 60 use cases.
However, currently, generative AI models mainly run in the cloud and require strong computing power to support them, making it difficult to popularize their applications. During the recent MWC2024 event, Meng Pu, Chairman of Qualcomm China, stated in an interview that if we want to enable more people to quickly and conveniently enjoy the new experiences brought by AI development, and to benefit more people with generative AI, we need to introduce AI technology to the places closest to users. Therefore, only by implementing generative AI functionality anytime and anywhere in terminal devices such as smartphones, mobile PCs, and connected vehicles that are within our reach, can we tightly integrate terminals with the cloud and promote the widespread application of generative AI.
At this year's MWC conference, Qualcomm launched a new Qualcomm AI Hub, providing developers with a comprehensive AI pre optimized model library that supports seamless deployment on terminals equipped with Snapdragon and Qualcomm platforms. At present, Qualcomm AI Hub can provide developers with over 75 mainstream AI and generative AI models, including familiar models such as Stable Diffusion, ControlNet, Baichuan-7B, etc. All of these models have been optimized to fully utilize the hardware acceleration capabilities of all cores (NPU, CPU, and GPU) within Qualcomm AI engine, thereby increasing inference speed by four times. Developers can quickly integrate these models seamlessly into applications, shorten product time to market, and leverage the many advantages of terminal side AI deployment, such as real-time, reliability, privacy, personalization, and cost advantages.
In order to serve more developers, Qualcomm has synchronously launched these optimization models on Qualcomm AI Hub, the world's largest code hosting platform GitHub, and the AI open source community Hugging Face. After registering and logging in, developers can run their own models on cloud hosting terminals equipped with the Qualcomm platform. They can also obtain new features and AI models in advance through the Qualcomm AI Hub, thereby accelerating product development and upgrading, and providing users with a brand new user experience. In addition, Qualcomm's AI Hub will continue to add new models and support more platforms and operating systems, continuously enriching the development ecosystem of generative AI applications and accelerating the widespread application of generative AI on various terminals.
Previously, Qualcomm had launched its first mobile platform specifically designed for generative AI - the third-generation Snapdragon 8, which supports a 10 billion parameter AI model running on smartphones. Meanwhile, Qualcomm has also launched the Snapdragon X Elite platform specifically for AI PCs, supporting generative AI models with over 13 billion parameters to run on the terminal side. At present, more than 10 smartphones equipped with the third-generation Snapdragon 8 have been launched, including the Xiaomi 14 series, iQOO12 series, Honor Magic6 series, Samsung Galaxy S24 series, OPPO Find X7 Ultra, etc. More AI PCs equipped with Snapdragon X Elite will also be launched later this year. All of these provide a broad hardware foundation for the application of generative AI on terminals, allowing more consumers to enjoy the innovative experience brought by generative AI anytime and anywhere.
Meng Pu further pointed out that with the improvement of chip computing power and continuous optimization of software, as well as the continuous development of AI models in generative AI applications, in the next 12 to 18 months, generative AI will be widely expanded from the current Snapdragon 8 series flagship phones to Snapdragon 7 series and even Snapdragon 6 series products, and consumers will also experience a richer variety of generative AI functions. Looking ahead to the future, Qualcomm will continue to be committed to expanding the commercial scale of terminal side AI in next-generation PCs, smartphones, software defined cars, XR devices, and the Internet of Things, making intelligent computing ubiquitous.
CandyLake.com 系信息发布平台,仅提供信息存储空间服务。
声明:该文观点仅代表作者本人,本文不代表CandyLake.com立场,且不构成建议,请谨慎对待。
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

日微牧 新手上路
  • 粉丝

    0

  • 关注

    0

  • 主题

    4