grow-ai-like-a-child/mechanical-reasoning

机械推理是一个多模态基准，旨在评估大型视觉语言模型中的机械和物理推理能力。该数据集包含148个样本，涉及图像和视频两种模态，以及选择题和判断题两种问题格式。数据集中的问题涵盖了多种物理推理挑战，如齿轮和滑轮系统、流体和浮力动力学、扭矩和平衡、输送机运动、机械优势、时间预测和因果关系等。

Mechanical Reasoning is a multimodal benchmark designed to evaluate mechanical and physical reasoning in large vision-language models (VLMs). The dataset contains 148 samples across two modalities: image and video, and two question formats: multiple-choice and true/false. The questions in the dataset span various physical reasoning challenges, including gear and pulley systems, fluid and buoyancy dynamics, torque and balance, conveyor motion, mechanical advantage, temporal prediction, and causality.