Full inetune mistral or llama model in GCP

[Mustafa21]

I am trying to full fintune llama2/mistral models on my costum data on GCP 
in full i mean not Lora/Qlora fintune with peft , 

i was searching for prebuild container but couldn't find one 
there is this   tutorial  here but its only fintune using lora / peft 

I need  to find a prebuild container for fintune these  models or a tutorial or explaination on how to do it. 

thanks