LM Studio AMD

https://wiki-fusion.win/index.php/AI_PC_Technology_Is_Changing_How_We_Work_%E2%80%94_Here%27s_What_That_Means

Running LM Studio on an AMD machine feels like a bit of a mixed bag—while it's fantastic that local LLMs can work with Vulkan and OpenCL backends, you might find yourself wrestling with driver compatibility or limited model support compared to Nvidia. The real challenge is that many optimized tools lean on CUDA, but with some tinkering you can still get models like Llama or Mistral running at a decent pace, especially with newer ROCm updates slowly bridging the gap.