Language-model training and adaptation · MLD-0333
DoRA
A LoRA-related method separating weight magnitude and direction during adaptation.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Improve low-rank adaptation flexibility.
Original caveat
It introduces additional parameters and implementation choices.
Context and aliases
Level: Intermediate. Stages: Training · Adaptation · Evaluation.
Tags: DL · SLM · LLM · NLP.
Aliases: Weight-decomposed low-rank adaptation.
Related concepts
Original references
- Transformers glossary ↗ · Consulted
- PEFT: LoRA conceptual guide ↗ · Consulted
- TRL: supervised fine-tuning ↗ · Reference link
- TRL documentation ↗ · Consulted
- Original reference URL ↗