It's in the name, Kubeflow is harder because it uses kubernetes which is much more complex than managing a conda environment. They also solve different problems and are used by different groups of people. ML flow really supports the development of and packaging of models, while Kubeflow is used orchestrate the infrastructure to support training and deploying models, which is a lot harder to do. You need dev ops because ultimately you weren't given sufficient permissions.
It's in the name, Kubeflow is harder because it uses kubernetes which is much more complex than managing a conda environment. They also solve different problems and are used by different groups of people. ML flow really supports the development of and packaging of models, while Kubeflow is used orchestrate the infrastructure to support training and deploying models, which is a lot harder to do. You need dev ops because ultimately you weren't given sufficient permissions.
Nice insight. Without knowing, you've just pretty much summarized this video 😅👍
Thanks! 🤝
Great video! Thanks
It was cioa cheese