Fine-tuning: Involves taking a pre-trained model and training it further on a labeled dataset specific to your domain Adapts the model to understand specific industry terminology, jargon, and requirements Uses supervised learning with labeled examples from the target domain Most common approach for customizing models for specific industries (legal, medical, finance, etc.) Maintains the general knowledge from pre-training while specializing for the specific use case

Fine tuning means taking a pre-trained model and training it further on labeled, domain-specific data so it learns specific terminology and requirements.

When a new domain specific task is added with the labeled data, Fine tuning comes into play. "Labeled Datasets" is a clear clue in this case, that's why it can't be pre-training technique.