LLM Optimization and Deployment on SiFive Intelligence

LLM Optimization and Deployment on SiFive Intelligence

Discover how RISC-V drives cutting-edge AI/ML technology.

Join us for an engaging webinar as we explore how SiFive is revolutionizing AI/ML deployment on RISC-V platforms. In this session, we'll dive into the SiFive AI/ML Software Stack, specifically designed for RISC-V, and demonstrate how to seamlessly deploy Pytorch Llama models to SiFive Intelligence platforms.

Discover the challenges of optimizing large language models (LLMs) and learn about our approach to achieving real-time Llama performance using the MLIR-based IREE compiler and runtime. This webinar is a must-attend for those interested in cutting-edge AI/ML solutions on RISC-V architectures.

Webinar Info

37 minutes

2024-10-16

Post Webinar Materials

Presentation Slides

2024-10-16 

View PDF

Speakers

Hong-Rong Hsu

Senior Staff Engineer

Hong-Rong Hsu has been leading and actively contributing to the Open Source Software team since 2019 and is currently serving as a Team Manager at SiFive.

In this role, he is responsible for both overseeing and hands-on development of AI/ML software, as well as managing AI/ML model e2e deployment.

Prior to joining SiFive, Hong-Rong gained valuable experience at MediaTek (2010–2018) and Bitmain (2018–2019).

His expertise spans AI/ML, system software, compilers, and RISC-V, with a strong focus on driving innovation and contributing to technical advancements in these areas.