Fine-tuning
Also called Finetuning
Fine-tuning continues training from an existing model to adapt it to data, a task, or a domain. It changes trainable parameters, rather than only changing the input supplied for one request.
[Hugging Face]In practice · hypothetical example
A team trains an existing model on labeled support examples, then tests it on a held-out set.
[Hugging Face]A little deeper
The starting point is a pretrained model. Training and evaluation data need separate roles so performance can be checked on examples not used for fitting. [Hugging Face]
A common mix-up
Adding a document to a prompt is fine-tuning.
Changing context for a request does not itself train model parameters. [Hugging Face]
Similar terms, different emphasis
Compare RAG & Fine-tuning →What distinguishes fine-tuning from adding prompt context?
Sources & editorial notes
Evidence: supported. Primary-source support for this scoped entry; publication approved by the project owner.
- Fine-tuning ↗ (opens in new tab)Hugging Face · Publication date unknown
Relevant section: Fine-tuning; Tokenization
Last editorial review: 2026-09-13 by project-owner.
First observed in this corpus: Unknown.
Revision history
Revision 2 · Created 2026-09-13 · Updated 2026-09-13
Project owner approved the current content for publication. Existing evidence scope and limitations remain applicable.