Glossary term
Glossary term
Foundations
A value that a model multiplies by another value. Training is the process of determining a model's ideal weights; inference is the process of using those learned weights to make predictions.
Click the icon to see an example of weights in a linear model.
See Linear regression in Machine Learning Crash Course for more information.
Created for this library
An ML engineer inspects weight distributions across layers to detect saturation in deep models.
A research engineer applies weight pruning to produce a smaller, faster production model.
An ML platform team versions model weights in a registry so audits can map outcomes back to the exact model that produced them.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License