VLM2Vec/MMEB-V2
收藏资源简介:
--- license: apache-2.0 task_categories: - visual-question-answering - video-classification language: - en viewer: false configs: - config_name: splits data_files: - split: eval path: - "video_tasks" - "image_tasks" --- # MMEB-V2 (Massive Multimodal Embedding Benchmark) Building upon on our original [**MMEB**](https://arxiv.org/abs/2410.05160), **MMEB-V2** expands the evaluation scope to include five new tasks: four video-based tasks — Video Retrieval, Moment Retrieval, Video Classification, and Video Question Answering — and one task focused on visual documents, Visual Document Retrieval. This comprehensive suite enables robust evaluation of multimodal embedding models across static, temporal, and structured visual data settings. **This Hugging Face repository contains only the raw image and video files used in MMEB-V2, which need to be downloaded in advance.** The test data for each task in MMEB-V2 is available [here](https://huggingface.co/VLM2Vec) and will be automatically downloaded and used by our code. More details on how to set it up are provided in the following sections. |[**Github**](https://github.com/TIGER-AI-Lab/VLM2Vec) | [**🏆Leaderboard**](https://huggingface.co/spaces/TIGER-Lab/MMEB) | [**📖MMEB-V2/VLM2Vec-V2 Paper (TBA)**](https://arxiv.org/abs/2410.05160) | | [**📖MMEB-V1/VLM2Vec-V1 Paper**](https://arxiv.org/abs/2410.05160) | ## 🚀 What's New - **\[2025.05\]** Initial release of MMEB-V2. ## Dataset Overview We present an overview of the MMEB-V2 dataset below: <img width="900" alt="abs" src="overview.png"> ## Dataset Structure The directory structure of this Hugging Face repository is shown below. For video tasks, we provide both sampled frames and raw videos (the latter will be released later). For image tasks, we provide the raw images. Files from each meta-task are zipped together, resulting in six files. For example, ``video_cls.tar.gz`` contains the sampled frames for the video classification task. ``` → video-tasks/ ├── frames/ │ ├── video_cls.tar.gz │ ├── video_qa.tar.gz │ ├── video_ret.tar.gz │ └── video_mret.tar.gz ├── raw videos/ (To be released) → image-tasks/ ├── mmeb_v1.tar.gz └── visdoc.tar.gz ``` After downloading and unzipping these files locally, you can organize them as shown below. (You may choose to use ``Git LFS`` or ``wget`` for downloading.) Then, simply specify the correct file path in the configuration file used by your code. ``` → MMEB ├── video-tasks/ │ └── frames/ │ ├── video_cls/ │ │ ├── UCF101/ │ │ │ └── video_1/ # video ID │ │ │ ├── frame1.png # frame from video_1 │ │ │ ├── frame2.png │ │ │ └── ... │ │ ├── HMDB51/ │ │ ├── Breakfast/ │ │ └── ... # other datasets from video classification category │ ├── video_qa/ │ │ └── ... # video QA datasets │ ├── video_ret/ │ │ └── ... # video retrieval datasets │ └── video_mret/ │ └── ... # moment retrieval datasets ├── image-tasks/ │ ├── mmeb_v1/ │ │ ├── OK-VQA/ │ │ │ ├── image1.png │ │ │ ├── image2.png │ │ │ └── ... │ │ ├── ImageNet-1K/ │ │ └── ... # other datasets from MMEB-V1 category │ └── visdoc/ │ └── ... # visual document retrieval datasets ```
license: apache-2.0 任务类别: - 视觉问答(visual-question-answering) - 视频分类(video-classification) 语言: - 英语 查看器:关闭 配置: - 配置名称:splits 数据文件: - 拆分:评估(eval) 路径: - "video_tasks" - "image_tasks" # MMEB-V2(大规模多模态嵌入基准测试集,Massive Multimodal Embedding Benchmark) 本数据集基于我们此前发布的[**MMEB**](https://arxiv.org/abs/2410.05160)构建,**MMEB-V2** 将评估范围拓展至五项全新任务:四项基于视频的任务——视频检索(Video Retrieval)、片段检索(Moment Retrieval)、视频分类(Video Classification)与视频问答(Video Question Answering),以及一项面向视觉文档的任务——视觉文档检索(Visual Document Retrieval)。这套全面的评估套件可实现多模态嵌入模型在静态、时序与结构化视觉数据场景下的稳健评估。 **本Hugging Face仓库仅包含MMEB-V2所使用的原始图像与视频文件,需提前完成下载。** MMEB-V2各任务的测试数据可通过[此处](https://huggingface.co/VLM2Vec)获取,我们的代码将自动完成下载并调用。后续章节将提供详细的配置步骤。 |[**Github仓库**](https://github.com/TIGER-AI-Lab/VLM2Vec) | [**🏆排行榜(Leaderboard)**](https://huggingface.co/spaces/TIGER-Lab/MMEB) | [**📖MMEB-V2/VLM2Vec-V2 论文(待更新,TBA)**](https://arxiv.org/abs/2410.05160) | | [**📖MMEB-V1/VLM2Vec-V1 论文**](https://arxiv.org/abs/2410.05160) | ## 🚀 更新内容 - **[2025.05]** MMEB-V2 正式首次发布。 ## 数据集概览 下文将展示MMEB-V2数据集的概览: <img width="900" alt="abs" src="overview.png"> ## 数据集结构 本Hugging Face仓库的目录结构如下所示。针对视频任务,我们同时提供采样帧与原始视频(后者后续将发布);针对图像任务,我们仅提供原始图像。每个元任务的文件均已打包压缩,共生成六个压缩包。例如,`video_cls.tar.gz` 包含视频分类任务的采样帧文件。 → video-tasks/ ├── frames/ │ ├── video_cls.tar.gz │ ├── video_qa.tar.gz │ ├── video_ret.tar.gz │ └── video_mret.tar.gz ├── raw videos/(待发布) → image-tasks/ ├── mmeb_v1.tar.gz └── visdoc.tar.gz 将这些文件下载至本地并解压后,可按照如下结构进行整理(可通过`Git LFS`或`wget`工具完成下载)。随后只需在代码所使用的配置文件中指定正确的文件路径即可。 → MMEB ├── video-tasks/ │ └── frames/ │ ├── video_cls/ │ │ ├── UCF101/ │ │ │ └── video_1/ # 视频ID │ │ │ ├── frame1.png # video_1的帧 │ │ │ ├── frame2.png │ │ │ └── ... │ │ ├── HMDB51/ │ │ ├── Breakfast/ │ │ └── ... # 视频分类类别下的其他数据集 │ ├── video_qa/ │ │ └── ... # 视频问答数据集 │ ├── video_ret/ │ │ └── ... # 视频检索数据集 │ └── video_mret/ │ └── ... # 片段检索数据集 ├── image-tasks/ │ ├── mmeb_v1/ │ │ ├── OK-VQA/ │ │ │ ├── image1.png │ │ │ ├── image2.png │ │ │ └── ... │ │ ├── ImageNet-1K/ │ │ └── ... # MMEB-V1类别下的其他数据集 │ └── visdoc/ │ └── ... # 视觉文档检索数据集




