MoE-Adapters4CL
论文代码：“通过混合专家适配器提升视觉语言模型的持续学习能力” CVPR2024。目录
摘要
Continual learning can empower vision-language models to continuously acquire new knowledge, without the need for access to the entire historical dataset. However, mitigating the performance degradation in large-scale models is non-trivial due to (i) parameter shifts throughout lifelong learning and (ii) significant computational burdens associated with full-model tuning. In this work, we present a parameter-efficient continual learning framework to alleviate long-term forgetting in incremental learning with vision-language models. Our approach involves the dynamic expansion of a pre-trained CLIP model, through the integration of Mixture-of-Experts (MoE) adapters in response to new tasks. To preserve the zero-shot recognition capability of vision-language models, we further introduce a Distribution Discriminative Auto-Selector (DDAS) that automatically routes in-distribution and out-of-distribution inputs to the MoE Adapter and the original CLIP, respectively. Through extensive experiments across various settings, our proposed method consistently outperforms previous state-of-the-art approaches while concurrently reducing parameter training burdens by 60%.方法
___
安装
conda create -n MoE_Adapters4CL python=3.9
conda activate MoE_Adapters4CL
conda install pytorch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 pytorch-cuda=11.8 -c pytorch -c nvidia
cd cil
pip install -r requirements.txt数据准备
目标数据集：Aircraft, Caltech101, CIFAR10, CIFAR100, DTD, EuroSAT, Flowers, Food, MNIST, OxfordPet, StanfordCars, SUN397, TinyImagenet。如果你在使用 Caltech101 时遇到问题，可以参考 issue#6。
更多详情可参考 ZSCL 的 datasets.md。非常感谢他们的出色工作！
模型权重
| | 模型 | 链接 | |------------------|----------------------------------------------------------------------|---------------------------------------------------------------------- | | full_shot_order1 | full_shot_order1_1000iters.pth | 百度网盘 / Google Drive | | few_shot_order1 | few_shot_order1_1000iters.pth | 百度网盘 / Google Drive |MTCL
测试阶段
示例：- 将权重文件移动到 MoE-Adapters4CL/ckpt
- ``
cd MoE-Adapters4CL/mtil` - 运行脚本 `
bash srcipts/test/Full_Shot_order1.sh`
训练阶段
示例：- 将权重文件移动到 MoE-Adapters4CL/ckpt
- `
cd MoE-Adapters4CL/mtil` - 运行脚本 `
bash srcipts/train/train_full_shot_router11_experts22_1000iters.sh`
类增量学习
训练阶段
示例：- `
cd cil` - `
bash run_cifar100-2-2.sh``
引用
@inproceedings{yu2024boosting,
title={Boosting continual learning of vision-language models via mixture-of-experts adapters},
author={Yu, Jiazuo and Zhuge, Yunzhi and Zhang, Lu and Hu, Ping and Wang, Dong and Lu, Huchuan and He, You},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={23219--23230},
year={2024}
}致谢
我们的代码库基于 wise-ft、Continual-CLIP 和 ZSCL 构建。我们感谢作者们分享他们的代码。--- Tranlated By Open Ai Tx | Last indexed: 2025-12-04 ---