GenAI finetuning explained

I am interested in understanding when to do what types of fine tuning. Example If I take a pretrained LLM and want to apply my business context to it so it can answer questions about my business specifically. Should I fine tune it on documents and specifications bout my business then perform instruction fine tuning?

Yeah you could do fine tuning on your business documents only, but maybe its better to do LoRA/Peft because the fine tuning might need a lot of resources.