Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the technical challenges and solutions for implementing AI applications on RISC-V platforms in this 40-minute conference talk from the Linux Foundation. Discover how to port, optimize, and run modern AI workloads including speech-to-text, LLM inference, and multimodal generation on resource-constrained RISC-V systems. Learn about practical implementations of popular open-source AI applications like Whisper, VLC with Speech-to-Text, webLLM, and text-to-image generation running entirely on a 64-bit 8-core RISC-V AI SoC. Examine the implications of DSP instruction emulation for RISC-V vector standard RVV 1.0, and understand how Vector DSP, Vulkan, and OpenCL backends can be leveraged for AI workloads. Gain insights into reproducible methodologies that enable the open-source community to extend AI capabilities on emerging RISC-V platforms, with practical guidance on how developers can contribute to or replicate these experimental implementations.
