Glossary

LoRA — Glossary

Low-Rank Adaptation: a parameter-efficient fine-tuning method that trains small adapter matrices instead of all model weights.

Also known as: Low-Rank Adaptation

Reviewed:

Full definition

LoRA (Hu et al.) freezes base model weights and learns low-rank updates injected into selected layers. This cuts memory and storage costs when specializing large models for a domain or format.

Adapters can be swapped or combined, enabling many specialized variants on one base. LoRA is still fine-tuning: you need curated data and regression evaluation.

For product decisions comparing adapters with retrieval, see RAG vs Fine-Tuning.

Related encyclopedia topic

Fine-Tuning — Fine-tuning adapts a pretrained model to a target task, domain, or style by continuing training on curated examples—often with parameter-efficient methods.

← Back to glossary · Learning hub