3 ms·
Yet, everyone misses reproducibility and data versioning :) So, talking about monitoring, training, and recording model drift is only a single side of the doma
by Fiahil 4y ago
Yet, everyone misses reproducibility and data versioning :)
So, talking about monitoring, training, and recording model drift is only a single side of the domain.
- KptMarchewa 4y ago> Yet, everyone misses reproducibility and data versioning :) Delta Lake/Apache Iceberg solves that.
- Fiahil 4y agoAbsolutely not. A single vendor/tech does not "solve" anything when the task at hand implies you need to entirely re-design data pipelines, ML modelling and benchmarking.
- alfalfasprout 4y agoLMAO no, no it doesn't and has major migration consequences for existing data warehouses. Reproducibility is more than just upstream data versioning.
- alfalfasprout 4y agoNot to mention dependency management! Since a lot of ML code is in Python this ends up being a very tricky thing to handle at scale (especially if you need to update dependencies, etc.)
- pharmakom 4y agoI didn’t care much for Docker until I started working in the Python ecosystem.