prar_shah
New Contributor III

Had the same question when I was trying to deploy my custom llm. How we did it was wrapping it inside a custom mlflow model and registering it.