Quick ML Pipeline using MLflow, Dagster, and Github Actions
This is a quick ML pipeline using:
MLflow for experiment tracking
Dagster for orchestration
Github for version control
Github Actions for CI
As I've been learning more about MLflow, I've been trying to figure out how to use it as part of a more standard software workflow (which often means tools like VCS, CI/CD, and orchestration).
From my naive perspective this all still boils down to: if I change something, can we minimize the chances that my changes either break our project or make it worse. And saying "hey, works on my machine and I can see the model is better" doesn't seem to cut it. As I layer in tools like MLflow, GH Actions, and Dagster, I'm getting more clarity.
Next up I want to build an actual dagster ML pipeline - my current setup is pitiful. Also, I want to see if I can surface some of the model metrics in Dagster (and maybe even the CI tool, GH Actions).
Смотрите видео Quick ML Pipeline using MLflow, Dagster, and Github Actions онлайн без регистрации, длительностью часов минут секунд в хорошем качестве. Это видео добавил пользователь Gus Cavanaugh 08 Декабрь 2022, не забудьте поделиться им ссылкой с друзьями и знакомыми, на нашем сайте его посмотрели 2,388 раз и оно понравилось 39 людям.