AMD launches Ryzen AI 1.0, which claims to help users build deployment models quickly.
Thanks to CTOnews.com netizen OC_Formula for the clue delivery! CTOnews.com, December 7 (Xinhua)-- AMD announced today that it will launch the Ryzen 8050 APU code-named Strix Point in 2024, powered by the XDNA 2 engine, with a focus on improving its ability to handle AI tasks.
According to the official AMD press release, AMD today also launched a Ryzen AI software that claims to help users easily build and deploy machine learning models on AI PC, and interested partners can click here to access product documentation.
According to the official introduction, users can use Ryzen AI software to build and deploy frameworks such as PyTorch and TensorFlow, and train models "on specific laptops supported by Ryzen AI".
CTOnews.com learned that Ryzen AI version 1.0 supports ONNX Runtime applications, and AMD claims that officials have provided a pre-optimized model repository on the Hugging Face platform, which claims to allow users to start and run AI models in minutes.
In addition, officials stress that laptops with Ryzen AI can uninstall AI models to NPU, releasing CPU, reducing power consumption, extending battery life, and building a variety of accessible AI applications:
With the launch of Ryzen AI software, developers can use Ryzen AI to build AI applications with advanced gesture recognition, biometric authentication and other auxiliary functions.
Developers can also get support for Automatic Speech Recognition (ASR) models (such as Whisper) and Large Language Models (LLM) (such as OPT and Llama-2) in advance to build a natural language voice interface for their applications and unlock other natural language processing features such as document summaries and e-mail assistance.
Related reading: "next year, AMD will improve the performance of Ruilong 8050 series APU:AI by three times, Zen 5 architecture CPU, RDNA 3 + kernel display"
