anabury/Cambridgedataset

剑桥A-Level多模态数据集是从剑桥A-Level过往试卷中提取的多模态数据集，包含文本、表格、图像、图表和数学方程。该数据集旨在用于训练和微调如Qwen-VL、LLaVA、MiniCPM-V、PaliGemma等多模态大型语言模型。数据集支持多模态推理、图表和图像解释、方程解析和符号推理、科学文本解释和逐步推理、OCR和表格理解、考试风格的问题回答训练。它非常适合教育领域的大型语言模型、STEM辅导和指导调优推理模型。

The Cambridge A-Level Multimodal Dataset is a multimodal dataset extracted from past papers of the Cambridge A-Level, containing text, tables, images, diagrams, and mathematical equations. It is designed for training and fine-tuning Multimodal Large Language Models (MLLMs) such as Qwen-VL, LLaVA, MiniCPM-V, PaliGemma, and others. The dataset enables multimodal reasoning, diagram and figure interpretation, equation parsing and symbolic reasoning, scientific text explanation and step-by-step reasoning, OCR and table understanding, and exam-style QA training. It is ideal for educational LLMs, STEM tutors, and instruction-tuned reasoning models.