argilla-io / ray-clay

Ray Clay is a tool to train and deploy models from Argilla using the Ray framework.

Commands

Server

Build

docker build --pull --rm -f "server.Dockerfile" -t ray-clay-server:latest "."

Run locally

poetry run uvicorn server.main:app

UI

Todo

Engineering [ ] create docker.compose [ ] model store - fsspec - BentoML?

UI: [ ] basic UI example VUE/Streamlit [ ] dashboard with running examples [ ] CRUD model [ ] CRUD training

Server: [ ] add ray train [ ] ray task for suggesstions [ ] task scheduler - kafka [ ] add more specific config for models