At Computex last month, AMD introduced the world's first block FP16 stable diffusion model, a breakthrough model that combines the accuracy of FP16 with performance of INT8. Now, this ongoing collaboration between AMD and Stability AI is proceeding to a new phase, with SD 3.0 Medium now available to download and run on Ryzen AI 300 or Ryzen AI MAX+ laptops.
AMD's Ryzen AI 300 series processors come with a built in Ryzen AI XDNA 2 NPU, which has enabled local AI on the go thanks to strong performance. The SD 3.0 Medium model delivers a significant increase in image quality and is optimised for the Ryzen AI XDNA 2 NPU, but you'll also need to ensure your system has at least 24GB to run it according to the minimum system requirements. If you meet those requirements today, then you can try out the latest improvements by downloading Amuse 3.1 by Tensorstack and running it in HQ mode. Once in HQ mode, you'll be able to toggle the ‘XDNA 2 NPU Stable Diffusion 3″ setting.
Stable Diffusion 3 Medium and will run on 24GB laptops while consuming only 9GB of memory. The use of this data type allows high quality, local AI image generation to be available on laptops with under 32GB of memory in a high precision format and without excessive quantization. Furthermore, a 2-stage pipeline powered by the AMD XDNA 2 NPU enhances the output from a 2MP image to a 4MP resolution.
Discuss on our Facebook page, HERE.
KitGuru Says: Do you use generative AI tools on your PC?