Skip to content

Model Training & Adaptation

LoRA

Low-rank adaptation

Adapting an AI by learning small additions rather than changing all of its existing learned settings.

Example

A team trains adapters for a specialized writing task.

Why people use it

It can adapt a large AI system while learning a relatively small addition.

What you'll hear

“Train the smaller addition for this task.”

What this means for you

Track adapter and base-AI system compatibility when deploying or sharing results.

Can you control it?

Developer-only

The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.

Common questions

Does a LoRA adapter contain an entire standalone base AI system?
Usually not. It typically depends on a compatible base AI system.
Can different additions serve different jobs?
Yes. Separate additions can be learned for different tasks, provided they fit the original system.
Can combining additions cause problems?
Yes. Their effects can interact, so combining them does not guarantee that each strength is preserved.

Related terms

Still have questions?

Up to 500 characters.

Ask LATHIC about AI. Relevant glossary entries may be included.

Your question, the glossary entries it matches, and a rotating pseudonymous identifier go to Microsoft Azure’s OpenAI service through Vercel AI Gateway to generate an answer. Zero retention and no training are required of the provider, and LATHIC does not save your question or answer. Privacy Notice