AMD on Thursday announced GAIA for “Generative AI Is Awesome” as a means of easily running local large language models (LLMs) on Ryzen AI PCs with the XDNA NPU. GAIA provides an easy software demonstration for Ryzen AI capabilities but sadly for the moment is limited to Microsoft Windows platforms.
Intel has their AI Playground and the like for their Windows customers while now AMD is rolling out GAIA as an easy software demonstration for running local, private LLMs on Ryzen AI hardware by using the NPU and/or integrated Radeon GPU. AMD GAIA can also run without any AMD accelerated support on any PC when switching to its ollama back-end. GAIA builds atop the ONNX TurnkeyML Lemonade SDK to run models like Llama and Phi.
This AMD blog post outlines the new GAIA project. GAIA is open-source under an MIT license and can be found on GitHub.
Sadly though at least for now it’s limited to Microsoft Windows 11. Though from the documentation with mentions like “Windows 11 Pro/Home (GAIA does not support macOS or Linux at this time, contact us if you need support for a particular platform)” it sounds like they may be open to supporting non-Windows platforms in the future… It would be nice to have a nice AMD software demonstration and benchmark for Ryzen AI on Linux now that Linux 6.14 introduces the Ryzen AI “AMDXDNA” accelerator driver and then on the user-space side users are encouraged to go with the IREE plug-in but without any convenient and compelling end-user software that’s really easy to immediately deploy for Ryzen AI Linux customers.
In any event it will be interesting to see what comes of this AMD GAIA project moving forward.
