How to Fine-Tune LLMs on AMD Strix Halo and Other Exotic AMD Hardware

After the first general general fine-tuning tutorial i posted here (https://www.promptinjection.net/p/the-ultimate-llm-ai-fine-tuning-guide-tutorial) some people asked if i can't make the same for AMD Strix Halo because approach here is quite different because of RoCM. https://preview.redd.it/sz5zy