MEFT: Memory-Efficient Fine-Tuning through Sparse Adapter

Jitai Hao, Weiwei Sun, Xin Xin, Qi Meng, Zhumin Chen, Pengjie Ren, Zhaochun Ren


Abstract
Parameter-Efficient Fine-tuning (PEFT) facilitates the fine-tuning of Large Language Models (LLMs) under limited resources. However, the fine-tuning performance with PEFT on complex, knowledge-intensive tasks is limited due to the constrained model capacity, which originates from the limited number of additional trainable parameters. To overcome this limitation, we introduce a novel mechanism that fine-tunes LLMs with adapters of larger size yet memory-efficient. This is achieved by leveraging the inherent activation sparsity in the Feed-Forward Networks (FFNs) of LLMs and utilizing the larger capacity of Central Processing Unit (CPU) memory compared to Graphics Processing Unit (GPU). We store and update the parameters of larger adapters on the CPU. Moreover, we employ a Mixture of Experts (MoE)-like architecture to mitigate unnecessary CPU computations and reduce the communication volume between the GPU and CPU. This is particularly beneficial over the limited bandwidth of PCI Express (PCIe). Our method can achieve fine-tuning results comparable to those obtained with larger memory capacities, even when operating under more limited resources such as a 24GB memory single GPU setup, with acceptable loss in training efficiency. Our codes are available at https://github.com/CURRENTF/MEFT.
Anthology ID:
2024.acl-long.129
Volume:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2375–2388
Language:
URL:
https://aclanthology.org/2024.acl-long.129
DOI:
10.18653/v1/2024.acl-long.129
Bibkey:
Cite (ACL):
Jitai Hao, Weiwei Sun, Xin Xin, Qi Meng, Zhumin Chen, Pengjie Ren, and Zhaochun Ren. 2024. MEFT: Memory-Efficient Fine-Tuning through Sparse Adapter. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2375–2388, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
MEFT: Memory-Efficient Fine-Tuning through Sparse Adapter (Hao et al., ACL 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/autopr/2024.acl-long.129.pdf