Add Inference Spec & CI

[AlejandroEsquivel]

Added:

• Added Inference Spec class to support Ray, Sagemaker, FastAPI in one go
• Added CI to support to deploy/update validator's inference endpoint on the cluster

Example Running FastAPI with Inference Spec:

pip install git+https://github.com/guardrails-ai/models-host@feat/adding-ray-setup
uvicorn models_host.fastapi:app

Example Running Ray Serve with Inference Spec:

# Refer to models-host repo for instructions if you don't have a GPU on the machine running ray serve
pip install git+https://github.com/guardrails-ai/models-host@feat/adding-ray-setup
serve run models_host.ray_serve:app