Run Wunjo with GPU on AMD

It’s no secret that most AI-driven applications require CUDA to fully unleash their potential, which means Nvidia GPUs are usually a must. Whether it’s Wunjo or countless other AI generation services, they all rely on Nvidia under the hood.

But something exciting happened yesterday—Wunjo successfully ran on an AMD GPU! You can find detailed instructions on how to make it work in Issue 65, with a follow-up in Issue 68 or updated README. In short, you’ll need to install ZLUDA, along with Torch and XForms from the RCom repo. Once set up, everything runs smoothly.

Now, for the downside: Nvidia has thrown a wrench in the works with a new license agreement for CUDA, effectively halting the progress of projects like ZLUDA. Even AMD has distanced itself from it.

So, what does this mean? Don't waste time! While ZLUDA is still functional, it’s the perfect opportunity to upgrade your AMD processor and install the software before it’s too late! Especially if you like to figure out the solution.