Run kohya-ss on RunPod to be able to finetune or train a LoRA

When your system isn't powerful enough to train or you don't want to stress it with that you can rent a GPU in the cloud. One such service is RunPod.

As you pay per minute it makes sense to try to minimize the setup time and maximize use the powerful system for training. As I couldn't find a good article that describes the setup in detail I've written one and posted it at https://github.com/StableLlama/kohya_on_RunPod

I'm happy about feedback and how to improve the article, either here or at GitHub.