LoRA

AI and Machine Learning · Efficiency and Deployment · 2021 · also: Low-rank adaptation · lora.yaml

Freezes the pretrained weights and learns a low-rank update beside them. Fine-tuning then costs a fraction of the memory, and the adapters can be swapped or merged afterwards.

Colour is the family; a dashed line is the second member of it.

Drag to pan · scroll to zoom · click a node to open it
G lora LoRA transformer Transformer lora->transformer adapts a frozen model through small low-rank updates

This node

References