Moving On From Azure

I have once again been stopped in my tracks, a common theme for this project. Currently there is a large demand for fine-tuning models on Azure. In light of this they have decided to indefinitely region lock the use of this service, and therefore there is no chance of fine-tuning a model on Azure. However, I have a backup plan! I am either going to go back to training a model using deepspeed or the new open sourced LLM OpenChatKit. https://github.com/togethercomputer/OpenChatKit

I have already made a singularity container for Deepspeed, but the fine-tuning of Openchatkit might prove to be easier. All I need to do is spin up an environment and I will be on my way to working with LLM again! Stay tuned for my update next week.

-Nick