Qibin Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Qibin Wang has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Bin Wang
MiLM Plus Xiaomi Inc Beijing China
1 shared publication Jian Luan
Xiaomi AI Lab
1 shared publicationWei Liu
Xiaomi AI Lab
1 shared publicationWeikai Xu
University of Electronic Science and Technology of China
1 shared publicationXiaolin Hu
Gaoling School of Artificial Intelligence Renmin University of China
1 shared publication
Latest Publications
- Open Access
PMSS: Pretrained Matrices Skeleton Selection for LLM Fine-tuning
Low-rank adaptation (LoRA) and its variants have recently gained much interest due to their ability to avoid excessive inference costs. However, LoRA still encounters the following challenges: (1) Limitation of low-rank assumption; and (2) Its initialization method may be suboptimal. To this end, we propose PMSS(Pre-trained Matrices Skeleton Selection), which enables high-rank updates with low costs while leveraging semantic and linguistic information inherent in pre-trained weight. It achieves this by selecting skeletons from the pre-trained weight matrix and only learning a small matrix instead. Experiments demonstrate that PMSS outperforms LoRA and other fine-tuning methods across tasks with much less trainable parameters. We demonstrate its effectiveness, especially in handling complex tasks such as DROP benchmark(+3.4%/+5.9% on LLaMA2-7B/13B) and math reasoning(+12.89%/+5.61%/+3.11% on LLaMA2-7B, Mistral-7B and Gemma-7B of GSM8K). The code and model will be released soon.
Open AccessPMSS: Pretrained Matrices Skeleton Selection for LLM Fine-tuning
Low-rank adaptation (LoRA) and its variants have recently gained much interest due to their ability to avoid excessive inference costs. However, LoRA still encounters the following challenges: (1) Limitation of low-rank assumption; and (2) Its initialization method may be suboptimal. To this end, we propose PMSS(Pre-trained Matrices Skeleton Selection), which enables high-rank updates with low costs while leveraging semantic and linguistic information inherent in pre-trained weight. It achieves this by selecting skeletons from the pre-trained weight matrix and only learning a small matrix instead. Experiments demonstrate that PMSS outperforms LoRA and other fine-tuning methods across tasks with much less trainable parameters. We demonstrate its effectiveness, especially in handling complex tasks such as DROP benchmark(+3.4%/+5.9% on LLaMA2-7B/13B) and math reasoning(+12.89%/+5.61%/+3.11% on LLaMA2-7B, Mistral-7B and Gemma-7B of GSM8K). The code and model will be released soon.