X-TAIL

#### Introduction of X-TAIL Benchmark

The X-TAIL (Cross-domain Task-Agnostic Incremental Learning) benchmark is composed of 10 datasets: Aircraft, Caltech101, DTD, EuroSAT, Flowers, Food, MNIST, OxfordPet, StanfordCars, and SUN397. It includes a total of 1,100 classes across all tasks.

We invite you to explore our work on the X-TAIL benchmark, published at ICML 2025: [LADA: Scalable Label-Specific CLIP Adapter for Continual Learning](https://icml.cc/virtual/2025/poster/43751). Code is available at: [https://github.com/MaolinLuo/LADA](https://github.com/MaolinLuo/LADA)

#### X-TAIL数据集介绍

X-TAIL(Cross-domain Task-Agnostic Incremental Learning)[1] 数据集由 Aircraft[2], Caltech101[3], DTD[4], EuroSAT[5], Flowers[6], Food[7], MNIST[8], OxfordPet[9], StanfordCars[10], and SUN397[11] 共10个数据集组成。所有任务包括总共 1,100 个类别。

欢迎关注我们在X-TAIL数据集上的发表于ICML2025的工作：[LADA: Scalable Label-Specific CLIP Adapter for Continual Learning](https://icml.cc/virtual/2025/poster/43751). 代码地址：[https://github.com/MaolinLuo/LADA](https://github.com/MaolinLuo/LADA)

#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

#### 引用

[1] Xu, Y., Chen, Y., Nie, J., Wang, Y., Zhuang, H., and Oku- mura, M. Advancing cross-domain discriminability in continual learning of vision-language models. Advances in neural information processing systems, 2024.

[2] Maji, S., Rahtu, E., Kannala, J., Blaschko, M., and Vedaldi, A. Fine-grained visual classification of aircraft. arXiv preprint arXiv:1306.5151, 2013.

[3] Fei-Fei, L., Fergus, R., and Perona, P. Learning generative visual models from few training examples: An incremen- tal bayesian approach tested on 101 object categories. In Conference on computer vision and pattern recognition workshop, 2004.

[4] Cimpoi, M., Maji, S., Kokkinos, I., Mohamed, S., and Vedaldi, A. Describing textures in the wild. In Pro- ceedings of the IEEE conference on computer vision and pattern recognition, 2014.

[5] Helber, P., Bischke, B., Dengel, A., and Borth, D. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of selected topics in applied earth observations and remote sensing, 2019.

[6] Nilsback, M.-E. and Zisserman, A. Automated flower clas- sification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & image processing, 2008.

[7] Bossard, L., Guillaumin, M., and Van Gool, L. Food-101– mining discriminative components with random forests. In European conference of computer vision, 2014.

[8] Deng, L. The mnist database of handwritten digit images for machine learning research [best of the web]. IEEE signal processing magazine, 2012.

[9] Parkhi, O. M., Vedaldi, A., Zisserman, A., and Jawahar, C. Cats and dogs. In 2012 IEEE conference on computer vision and pattern recognition, 2012.

[10] Krause, J., Stark, M., Deng, J., and Fei-Fei, L. 3d object rep- resentations for fine-grained categorization. In Proceed- ings of the IEEE international conference on computer vision workshops, 2013.

[11] Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., and Torralba, A. Sun database: Large-scale scene recognition from abbey to zoo. In IEEE computer society conference on computer vision and pattern recognition, 2010.

#### X-TAIL基准数据集介绍
X-TAIL（跨域任务无关增量学习，Cross-domain Task-Agnostic Incremental Learning）基准数据集由10个数据集组成，分别为飞行器细分类数据集（Aircraft）、加州理工101类物体数据集（Caltech101）、野外纹理描述数据集（DTD）、欧洲卫星土地利用与覆盖分类数据集（EuroSAT）、花卉分类数据集（Flowers）、食品分类数据集（Food）、手写数字数据集（MNIST）、牛津宠物数据集（OxfordPet）、斯坦福汽车细分类数据集（StanfordCars）以及大规模场景分类数据集（SUN397）。全任务总计涵盖1100个类别。

诚邀您关注我们发表于ICML 2025国际机器学习会议的相关研究成果：《LADA：面向持续学习的可扩展标签专属CLIP适配器》（LADA: Scalable Label-Specific CLIP Adapter for Continual Learning），链接为：[https://icml.cc/virtual/2025/poster/43751](https://icml.cc/virtual/2025/poster/43751)。代码开源地址为：[https://github.com/MaolinLuo/LADA](https://github.com/MaolinLuo/LADA)

#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

#### 引用
[1] Xu, Y., Chen, Y., Nie, J., Wang, Y., Zhuang, H., 以及 Okumura, M. 《提升视觉语言模型持续学习中的跨域判别能力（Advancing cross-domain discriminability in continual learning of vision-language models）》. 《神经信息处理系统进展》, 2024.
[2] Maji, S., Rahtu, E., Kannala, J., Blaschko, M., 以及 Vedaldi, A. 《飞行器细分类视觉识别（Fine-grained visual classification of aircraft）》. arXiv预印本arXiv:1306.5151, 2013.
[3] Fei-Fei, L., Fergus, R., 以及 Perona, P. 《基于少量训练样本的生成式视觉模型学习：在101类物体数据集上验证的增量贝叶斯方法（Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories）》. 2004年IEEE计算机视觉与模式识别研讨会会议论文集, 2004.
[4] Cimpoi, M., Maji, S., Kokkinos, I., Mohamed, S., 以及 Vedaldi, A. 《野外纹理描述（Describing textures in the wild）》. IEEE计算机视觉与模式识别会议论文集, 2014.
[5] Helber, P., Bischke, B., Dengel, A., 以及 Borth, D. 《EuroSAT：面向土地利用与土地覆盖分类的新型数据集与深度学习基准（Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification）》. IEEE应用地球观测与遥感精选主题期刊, 2019.
[6] Nilsback, M.-E. 以及 Zisserman, A. 《大规模类别下的自动化花卉分类（Automated flower classification over a large number of classes）》. 2008年第六届印度计算机视觉、图形与图像处理会议, 2008.
[7] Bossard, L., Guillaumin, M., 以及 Van Gool, L. 《Food-101——利用随机森林挖掘判别性组件（Food-101– mining discriminative components with random forests）》. 欧洲计算机视觉会议, 2014.
[8] Deng, L. 《MNIST：面向机器学习研究的手写数字图像数据集（The mnist database of handwritten digit images for machine learning research [best of the web]）》. IEEE信号处理杂志, 2012.
[9] Parkhi, O. M., Vedaldi, A., Zisserman, A., 以及 Jawahar, C. 《猫与狗（Cats and dogs）》. 2012年IEEE计算机视觉与模式识别会议, 2012.
[10] Krause, J., Stark, M., Deng, J., 以及 Fei-Fei, L. 《面向细分类识别的3D物体表征（3d object representations for fine-grained categorization）》. IEEE国际计算机视觉研讨会论文集, 2013.
[11] Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., 以及 Torralba, A. 《SUN数据库：从修道院到动物园的大规模场景识别（Sun database: Large-scale scene recognition from abbey to zoo）》. IEEE计算机学会计算机视觉与模式识别会议, 2010.