Artificial intelligence (AI) continues to be at the forefront of operations driven by companies large, medium and small. In this case it is Qualcomm has reinforced its commitment to providing buildable AI technology that is accessible and customizable to each developer’s needs. To do this, he feeds his AI Hub platform cheater New models. During the 2024 edition of Mobile World Congress (MWC), the San Diego-based multinational company announced a library of more than 75 pre-optimized AI models that will be implemented on enterprise devices with Snapdragon platforms.
According to the information provided thenThe above models take advantage of acceleration Hardware across all cores in the AI engine From Qualcomm. Thanks to this ability they have achieved Delivering results with „four times faster” inference times. This way, developers can integrate them into their applications to take advantage of AI on the device. One of the main benefits gained Immediacy, privacy and personalization, as well as cost savings in applications.
Currently, Qualcomm AI Hub has a total of 95 AI models, including Standard Diffusion, Whisper, ConvNext-Tiny, AOT-GAN, ControlNet and Baichuan 7B. Corporation assures that all of them will be optimized Provides better performance by using less memory and improving energy efficiency.