agent-studio/IDM-Multiple

IDM-Multiple是IDMBench的一部分，旨在评估在没有动作标签的互联网规模视频中进行学习的能力。该数据集要求视觉语言模型(VLMs)在屏幕录制视频片段中回答动作序列。请注意，这是一个测试集，不应用于训练。

IDM-Multiple is part of the IDMBench for evaluating the ability to learn from Internet-scale videos without action labels. Specifically, this dataset requires Visual Language Models (VLMs) to answer an action sequence within a video clip of screen recording. Please note that this dataset is a test set and should not be used for training.