Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Model Stock: All we need is just a few fine-tuned models

About

This paper introduces an efficient fine-tuning method for large pre-trained models, offering strong in-distribution (ID) and out-of-distribution (OOD) performance. Breaking away from traditional practices that need a multitude of fine-tuned models for averaging, our approach employs significantly fewer models to achieve final weights yet yield superior accuracy. Drawing from key insights in the weight space of fine-tuned weights, we uncover a strong link between the performance and proximity to the center of weight space. Based on this, we introduce a method that approximates a center-close weight using only two fine-tuned models, applicable during or after training. Our innovative layer-wise weight averaging technique surpasses state-of-the-art model methods such as Model Soup, utilizing only two fine-tuned models. This strategy can be aptly coined Model Stock, highlighting its reliance on selecting a minimal number of models to draw a more optimized-averaged model. We demonstrate the efficacy of Model Stock with fine-tuned models based upon pre-trained CLIP architectures, achieving remarkable performance on both ID and OOD tasks on the standard benchmarks, all while barely bringing extra computational demands. Our code and pre-trained models are available at https://github.com/naver-ai/model-stock.

Dong-Hwan Jang, Sangdoo Yun, Dongyoon Han• 2024

Related benchmarks

TaskDatasetResultRank
Mathematical ReasoningGSM8K
Accuracy59.67
983
Mathematical ReasoningMATH
Accuracy16.64
643
Multiple-choice Question AnsweringMMLU-Pro
MMLU-Pro Overall Accuracy36.8
116
Safety AlignmentHarmBench
ASR17.25
88
Code GeneratingMBPP
Pass@147.8
88
Multiple-choice Question AnsweringSciQ
Accuracy95.2
74
Safety AlignmentSORRY-Bench
ASR12.67
40
Mathematical ReasoningGSM8K Platinum
Accuracy59
37
Multilingual Mathematical ReasoningMSVAMP
Accuracy (English)32.3
33
Code GeneratingHumanEvalPack
Pass@139.02
24
Showing 10 of 12 rows

Other info

Follow for update