Azure API for LLM changes depending on what datacenter you are calling. It is bonkers. In fact it is so bad that at work we are hosting our own LLMs on azure GP

Azure API for LLM changes depending on what datacenter you are calling. It is bonkers. In fact it is so bad that at work we are hosting our own LLMs on azure GPU machines rather than use their API. (Which means we only have small models at much higher cost…)