It assumes you have an account on VAST-AI and understand what I'm talking about, so go there, create an account, and look around. (Skip this step if your local GPU has 24 GB VRAM, like an RTX 4090)
The notebook is "llama3_8b_finetune_own_data.ipynb"