I think it will be possible. On linux it works on AMD and there is a plugin for the auto1111 ui which supports amd gpus, apparently. The latter one is harder to install, though. :/
But I really cannot tell how long it'll take until we get an AMD gpu version. I thought it would have been done much sooner, but there's now this plugin for another UI... so - as far as I know - it might be just a few weeks away or several months. :/
On my 10 year old FX 8350 cpu, it took 8 minutes to generate a 512x512 image. On yours maybe 3-4 minutes, but I don't know.
So it won't be super fun, but if you take a good model/prompt from civitai and leave it running while doing something else, it miiiight be okayish.
Viewing post in InvokeAI - The Stable Diffusion Toolkit comments
You won't have to uninstall windows. :3
Though, this standalone only works with windows. :( So for Linux you have to follow the automatic installer guide: https://invoke-ai.github.io/InvokeAI/installation/010_INSTALL_AUTOMATED/#walk-th... and the rocm (= "AMD cuda") guide https://invoke-ai.github.io/InvokeAI/installation/030_INSTALL_CUDA_AND_ROCM/#wal...
It's not crazy hard, but it's obviously a bit trickier. :)