TIGER-Lab/VisPhyWorld-Sub-Generated-Videos

该数据集包含为VisPhyBench/VisPhyWorld的sub分割生成的视频，按渲染引擎和模型组织。它主要用于模型比较、错误分析以及代码驱动或直接视频生成结果的定性检查。数据集包含来自threejs、p5js和video引擎的不同模型生成的视频文件，如gpt-5、claude-sonnet-4-5、gemini-3-pro、gpt-4.1、qwen3-vl-plus、svd-img2vid、sora-2和veo31等。数据集是VisPhyWorld集合的配套资源，用于定性评估。

This dataset contains generated videos for the `sub` split of **VisPhyBench / VisPhyWorld**, organized by rendering engine and model. It is intended for model comparison, error analysis, and qualitative inspection of code-driven or direct video generation results. The dataset includes video files generated by different models from threejs, p5js, and video engines, such as gpt-5, claude-sonnet-4-5, gemini-3-pro, gpt-4.1, qwen3-vl-plus, svd-img2vid, sora-2, and veo31. The dataset serves as a companion asset for the VisPhyWorld collection and is intended for qualitative evaluation.