Glossary term
Glossary term
Training and Fine-Tuning
The process of determining the best model.
Created for this library
An ML platform team treats model training as a reproducible pipeline with versioned datasets, code, and configurations.
A research team logs model training metrics and artifacts to a tracking server so experiments can be compared and resumed.
An ML platform team schedules model training jobs nightly so production models are always built on the freshest approved data.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License