LoRA (Low-Rank Adaptation)

A cheap fine-tuning technique that updates only a small number of parameters — used everywhere in 2026.

Definition

LoRA fine-tunes a model by adding small "adapter" layers and only training those. Updates ~1% of parameters but achieves close-to-full-finetune quality. Most contractor-data gigs feed LoRA-based training.

Examples

  • LoRA fine-tuning Llama on a specialty dataset
  • Stable Diffusion LoRAs for art styles

Related terms

Looking for actual gigs that use LoRA (Low-Rank Adaptation)? Browse all live listings or filter by $100+/hr.