Llama 3.1 finetuning on your own data and infra, without vendor lock-in | Dark Hacker News