Experiment tracking for multiple ML models using mlflow in a single main evaluation

Could you, in your experience, show an article or an experiment tracking example and only version “Multi-independent models, but one input-> multiple models-> one output” to get a single main score and conveniently compare sub-scores? see an example project in the diagram:

I understand and tried to use W&B, MLFlow, DVC, Neptune. ai, DagsHub for only one model, but I’m not sure one is convenient to use for multi-independent models. I also did not find it in Google for the approximate phrase “ML tracking experiment and management for multi models”

Hi @yayay , thank you for the question. As you intend to train multiple models at once, our recommended approach to this is utilize multiple projects where for each, each models experimental is tracked. You could track the same model within the projects using artifacts. See our document for model versioning tips. Please let me know if you have any questions.

1 Like

Hi @mohammadbakir , thanks for the answer. I will carefully read, do and still answer you:)

Thank you for the update @yayay , I will mark this resolved for now but please do re open the conversation if you need to.