Models Capabilities Use Cases Benchmarks Papers Glossary

Models Capabilities Use Cases Benchmarks Papers Glossary

About Privacy Terms RSS

ThinkLLM

Glossary/DoRA (Weight-Decomposed Low-Rank Adaptation)

DoRA (Weight-Decomposed Low-Rank Adaptation)

techniques

A fine-tuning method that adapts model weights by separately learning magnitude and direction changes, extending LoRA.

DoRA (Weight-Decomposed Low-Rank Adaptation) — Glossary — ThinkLLM