[D] Fine Tuning Code LLama2 on new programing language

I am exploring whether fine-tuning works for teaching a new programming language to Code LLama2 that it was not trained on, for example, Ruby. Has anyone tried this?

I tried few-shot prompting. Code LLama seems to be doing a fair job. Not sure if anyone has done full-blown fine-tuning on any new languages and want to hear the success rate.

code code llama example fair few-shot fine-tuning job language language i llama llama2 machinelearning new programming language programing programming programming language prompting ruby teaching