LoRA: Low-Rank Adaptation of Large Language Models
Source
Edward J. Hu et al.
Author
Edward J. Hu et al.
Published
Terms in this piece · Glossary
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
fine-tuning — Taking a trained model and training it a bit more on your own examples so it gets better at one specific job.
Why it matters
It is why fine-tuningTaking a trained model and training it a bit more on your own examples so it gets better at one specific job.Full definition → is something you can do on your own hardware. Freezing the base model and training low-rank adapters cuts trainable parameters by orders of magnitude, adds no latency once merged, and lets many task-specific adapters share one base model.
