scaling laws

Empirical observations in machine learning that describe how the performance of models improves as a function of increased data, model size, or computational resources. Understanding scaling laws helps predict how enhancements will affect model training.

19 papers