QLoRA: Efficient Finetuning of Quantized LLMs
We present QLoRA, an efficient finetuning approach that reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU while preserving full 16-bit finetuning task performance. QLoRA backpropagates gradients through a frozen, 4-bit quantized pretrained language model into Low Ran...
Gespeichert in:
Hauptverfasser: | , , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | We present QLoRA, an efficient finetuning approach that reduces memory usage
enough to finetune a 65B parameter model on a single 48GB GPU while preserving
full 16-bit finetuning task performance. QLoRA backpropagates gradients through
a frozen, 4-bit quantized pretrained language model into Low Rank
Adapters~(LoRA). Our best model family, which we name Guanaco, outperforms all
previous openly released models on the Vicuna benchmark, reaching 99.3% of the
performance level of ChatGPT while only requiring 24 hours of finetuning on a
single GPU. QLoRA introduces a number of innovations to save memory without
sacrificing performance: (a) 4-bit NormalFloat (NF4), a new data type that is
information theoretically optimal for normally distributed weights (b) double
quantization to reduce the average memory footprint by quantizing the
quantization constants, and (c) paged optimziers to manage memory spikes. We
use QLoRA to finetune more than 1,000 models, providing a detailed analysis of
instruction following and chatbot performance across 8 instruction datasets,
multiple model types (LLaMA, T5), and model scales that would be infeasible to
run with regular finetuning (e.g. 33B and 65B parameter models). Our results
show that QLoRA finetuning on a small high-quality dataset leads to
state-of-the-art results, even when using smaller models than the previous
SoTA. We provide a detailed analysis of chatbot performance based on both human
and GPT-4 evaluations showing that GPT-4 evaluations are a cheap and reasonable
alternative to human evaluation. Furthermore, we find that current chatbot
benchmarks are not trustworthy to accurately evaluate the performance levels of
chatbots. A lemon-picked analysis demonstrates where Guanaco fails compared to
ChatGPT. We release all of our models and code, including CUDA kernels for
4-bit training. |
---|---|
DOI: | 10.48550/arxiv.2305.14314 |