Compared to MLFlow / ModelDB / Polyaxon

[elgalu]

How does this project compares to MLFlow / ModelDB / Polyaxon?

Is it better support for real-time? or multi-tenancy? something else?

Thanks in advance.

[samzer]

Hey Leo,

Compared to Mlflow and ModelDB, Modelchimp has:

• Realtime: Realtime status of the currently running experiment and real-time epoch metric graphs
• Storage of assets with meta information: Useful post forensic prediction on images and text where the metric associated with each prediction is stored
• Collaboration: Define a single project and have multiple team members track the experiments at a single place
• Code Snapshot: Code used for the particular experiment. ( Mlflow stores the current git commit id )

ModelDB is more focussed on machine learning algorithms. They have formed a company around it called Verta.ai and they'll be releasing a V2 version which will have support for deep learning.

I wasn't aware of Polyaxon and looking at their webpage, it seems really cool. It looks like the most holistic solution out there.

Hope this answer your questions.

[elgalu]

Ping @mouradmourafiq from Polyaxon; you have been praised:)

[elgalu]

thanks!!