Train Small, Infer Large: Memory-Efficient LoRA Training for Large Language Models

计算机科学 培训(气象学) 自然语言处理 人工智能 地理 气象学
作者
Jun Zhang,Jue Wang,Huan Li,Lidan Shou,Ke Chen,Yang You,Guiming Xie,Xuejian Gong,Kaijun Zhou
出处
期刊:Cornell University - arXiv
标识
DOI:10.48550/arxiv.2502.13533
摘要

Large Language Models (LLMs) have significantly advanced natural language processing with exceptional task generalization capabilities. Low-Rank Adaption (LoRA) offers a cost-effective fine-tuning solution, freezing the original model parameters and training only lightweight, low-rank adapter matrices. However, the memory footprint of LoRA is largely dominated by the original model parameters. To mitigate this, we propose LoRAM, a memory-efficient LoRA training scheme founded on the intuition that many neurons in over-parameterized LLMs have low training utility but are essential for inference. LoRAM presents a unique twist: it trains on a pruned (small) model to obtain pruned low-rank matrices, which are then recovered and utilized with the original (large) model for inference. Additionally, minimal-cost continual pre-training, performed by the model publishers in advance, aligns the knowledge discrepancy between pruned and original models. Our extensive experiments demonstrate the efficacy of LoRAM across various pruning strategies and downstream tasks. For a model with 70 billion parameters, LoRAM enables training on a GPU with only 20G HBM, replacing an A100-80G GPU for LoRA training and 15 GPUs for full fine-tuning. Specifically, QLoRAM implemented by structured pruning combined with 4-bit quantization, for LLaMA-3.1-70B (LLaMA-2-70B), reduces the parameter storage cost that dominates the memory usage in low-rank matrix training by 15.81$\times$ (16.95$\times$), while achieving dominant performance gains over both the original LLaMA-3.1-70B (LLaMA-2-70B) and LoRA-trained LLaMA-3.1-8B (LLaMA-2-13B).
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
balabala3完成签到,获得积分10
1秒前
2秒前
xr完成签到,获得积分10
3秒前
念姬发布了新的文献求助10
4秒前
星光发布了新的文献求助10
5秒前
董又又又又完成签到,获得积分10
6秒前
6秒前
Leexxxhaoo完成签到,获得积分10
6秒前
勤劳莹芝发布了新的文献求助10
7秒前
有魅力的寻梅完成签到,获得积分20
8秒前
阿成完成签到,获得积分10
8秒前
8秒前
叼面包的数学狗完成签到 ,获得积分10
11秒前
11秒前
12秒前
橙子大王发布了新的文献求助10
12秒前
YY完成签到,获得积分10
12秒前
阿成发布了新的文献求助10
12秒前
12秒前
阿莲呐发布了新的文献求助20
12秒前
14秒前
JamesPei应助XXXX采纳,获得10
15秒前
16秒前
16秒前
坦率班完成签到 ,获得积分10
17秒前
星河发布了新的文献求助20
17秒前
七七完成签到,获得积分10
18秒前
19秒前
SLY完成签到 ,获得积分10
20秒前
20秒前
所所应助跳跳虎采纳,获得10
20秒前
wanci应助seedcode采纳,获得10
21秒前
我是老大应助完犊子采纳,获得10
21秒前
Kevin发布了新的文献求助30
21秒前
22秒前
23秒前
灰灰灰完成签到,获得积分10
25秒前
牛牛眉目发布了新的文献求助10
25秒前
27秒前
哈哈哈完成签到,获得积分10
27秒前
高分求助中
A new approach to the extrapolation of accelerated life test data 1000
Cognitive Neuroscience: The Biology of the Mind 1000
Technical Brochure TB 814: LPIT applications in HV gas insulated switchgear 1000
Immigrant Incorporation in East Asian Democracies 500
Nucleophilic substitution in azasydnone-modified dinitroanisoles 500
不知道标题是什么 500
A Preliminary Study on Correlation Between Independent Components of Facial Thermal Images and Subjective Assessment of Chronic Stress 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 遗传学 基因 物理化学 催化作用 冶金 细胞生物学 免疫学
热门帖子
关注 科研通微信公众号,转发送积分 3966370
求助须知:如何正确求助?哪些是违规求助? 3511789
关于积分的说明 11159900
捐赠科研通 3246400
什么是DOI,文献DOI怎么找? 1793416
邀请新用户注册赠送积分活动 874427
科研通“疑难数据库(出版商)”最低求助积分说明 804388