JessicaE/OpenSeeSimE-Fluid-Mini

OpenSeeSimE-Fluid-Mini是cmudrc/OpenSeeSimE-Fluid数据集的1%分层子集，旨在减少计算资源的同时评估视觉语言模型。数据集包含1040个样本，覆盖了五种源文件（Bent Pipe、Converging Nozzle、Heat Exchanger、Heat Sink、Mixing Pipe）、三种问题类型（Binary、Multiple Choice、Spatial）和两种媒体类型（image、video）。每个样本包含文件名、源文件、问题文本、问题类型、问题ID、答案、答案选项、正确答案索引、图像或视频等特征。数据集适用于工程模拟问答的基准评估、评估管道的烟雾测试以及在存储或带宽受限的情况下的比较研究。

OpenSeeSimE-Fluid-Mini is a stratified 1% subset of the cmudrc/OpenSeeSimE-Fluid dataset, designed to evaluate vision-language models at a reduced compute footprint. The dataset contains 1,040 samples covering five source files (Bent Pipe, Converging Nozzle, Heat Exchanger, Heat Sink, Mixing Pipe), three question types (Binary, Multiple Choice, Spatial), and two media types (image, video). Each sample includes features such as file name, source file, question text, question type, question ID, answer, answer choices, correct choice index, image or video. The dataset is intended for benchmark evaluation of vision-language models on engineering simulation question answering, smoke-testing of evaluation pipelines, and comparative studies under storage or bandwidth constraints.