Dettmers and colleagues: QLoRA quantized adapter fine-tuning

What changed

Fine-tuned low-rank adapters through frozen four-bit model weights, combining NormalFloat storage, double quantization and paged optimizers to reduce memory use.

Scientific scope

Fine-tuned low-rank adapters through frozen four-bit model weights, combining NormalFloat storage, double quantization and paged optimizers to reduce memory use. Quantized storage is dequantized for higher-precision computation. Quality retention and memory figures are empirical for the studied tasks; QLoRA builds on the already catalogued LoRA method.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Quantized storage is dequantized for higher-precision computation. Quality retention and memory figures are empirical for the studied tasks; QLoRA builds on the already catalogued LoRA method.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• QLoRA: Efficient Finetuning of Quantized LLMs
• NeurIPS 2023: QLoRA proceedings record

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction