Definition

Pre-trained models are machine learning models that have already been trained on a large dataset for a particular task. They can then be fine-tuned for more specific applications.

Why it matters (in Poovi’s context)

Discussed as a foundational element in AI development, leading to new job roles and applications, particularly in specialised domains.

Key properties or components

  • Large datasets
  • Fine-tuning
  • Transfer learning
  • Foundation models

Contradictions or debates

None.

Sources