transfer learning

A better approach is to pre-train the model on a larger dataset first, then fine-tune it on this target dataset (classically, we call this approach “transfer learning”).

Language Models for Text Classification: From Bag-of-Words to Jev Sebastian Raschka, PhD 2026