DataOrigin/k12-video-solutions-india

--- license: other task_categories: - video-classification - question-answering language: - hi - en - ta - te - ml - bn - mr - gu - or - pa - as tags: - education, - k12, - india, - multilingual, - mathematics, - science, - curriculum, - indic, pretty_name: K-12 Video Solutions India size_categories: - 1M<n<10M --- # K-12 Video Solutions India ## Dataset Description A large-scale collection of step-by-step video solutions to K-12 curriculum questions, solved by subject matter experts. Produced by Prepp, India's largest K-12 learning platform, operated by Collegedunia Web Private Limited. ## Dataset Summary - **Total videos:** 3.8 million (38 lakh) - **Content type:** Step-by-step question-answer video solutions - **Subjects:** Mathematics, Science, Social Studies, English, Regional Languages - **Grades:** Class 1 through Class 12 - **Boards:** 33 Indian educational boards including CBSE, ICSE, and all major state boards - **Languages:** Hindi, English, Tamil, Telugu, Kannada, Malayalam, Marathi, Bengali, Gujarati, Odia, Punjabi, Assamese - **Format:** Video (MP4) with audio narration and visual problem-solving ## Sample Data Three sample videos are available in this repository demonstrating: - Sample 1: Mathematics problem solving (Class 10, CBSE) - Sample 2: Science concept explanation (Class 8, CBSE) - Sample 3: Regional language content (Hindi medium) ## Key Features - **Human-verified:** All solutions reviewed by qualified subject experts - **Curriculum-mapped:** Each video tagged to specific board, grade, subject, chapter, and topic - **Multimodal:** Rich combination of audio narration, handwritten working, diagrams, and step-by-step visual explanation - **Low-resource languages:** Includes Odia, Assamese, and Punjabi — genuinely low-resource for AI training globally ## Intended Uses - Training multimodal AI models for educational reasoning - STEM problem-solving model development - Indic language speech and audio model training - Video understanding and question-answering model fine-tuning - Multilingual educational AI applications ## Data Collection and Rights All content is proprietary, produced by Prepp's in-house content team and contracted subject matter experts under work-for-hire agreements. Content is ethically sourced and curriculum-mapped. Full dataset licensing is available for commercial AI training purposes. ## Licensing and Commercial Access This repository contains sample data only. The full dataset of 3.8 million videos is available for commercial AI training licensing. **For licensing inquiries contact:** Ankit Dubey — Head of AI Data Partnerships, Collegedunia [e-mail] ## Dataset Curator [Collegedunia Web Private Limited](https://collegedunia.com) | [Prepp](https://prepp.in) Gurugram, Haryana, India许可证：其他 任务类别： - 视频分类 - 问答 语言： - 印地语 - 英语 - 泰米尔语 - 泰卢固语 - 马拉雅拉姆语 - 孟加拉语 - 马拉地语 - 古吉拉特语 - 奥里亚语 - 旁遮普语 - 阿萨姆语 标签： - 教育 - K12 - 印度 - 多语言 - 数学 - 科学 - 课程体系 - 印度语系 展示名：印度K-12课程视频解答集 规模类别：100万<样本数<1000万 # 印度K-12课程视频解答集 ## 数据集说明 本数据集为大规模分步式K-12课程习题视频解答合集，所有解答均由学科专家完成。该数据集由印度最大的K-12学习平台Prepp出品，由Collegedunia Web Private Limited运营。 ## 数据集概览 - **总视频数：380万（38 lakh，印度计数体系单位）** - **内容类型：分步式问答视频解答** - **涵盖学科：数学、科学、社会研究、英语及各区域语言** - **适用年级：1年级至12年级** - **适配考试委员会：涵盖33个印度教育委员会，包括中央中等教育委员会（CBSE）、印度中等教育证书考试（ICSE）及所有主要邦级教育委员会** - **支持语言：印地语、英语、泰米尔语、泰卢固语、卡纳达语、马拉雅拉姆语、马拉地语、孟加拉语、古吉拉特语、奥里亚语、旁遮普语、阿萨姆语** - **格式：带语音讲解与可视化解题过程的MP4视频** ## 示例数据 本仓库提供三个示例视频，分别展示： - 示例1：数学题解答（10年级，CBSE） - 示例2：科学概念讲解（8年级，CBSE） - 示例3：区域语言内容（印地语授课模式） ## 核心特性 - **人工审核：所有解答均由具备资质的学科专家审核** - **课程匹配：每个视频均标注对应考试委员会、年级、学科、章节与知识点** - **多模态特性：融合语音讲解、手写演算、图表与分步可视化讲解的丰富内容** - **低资源语言支持：涵盖奥里亚语、阿萨姆语与旁遮普语，这类语言在全球AI训练领域均属于低资源语言** ## 预期用途 - 训练面向教育推理的多模态AI模型 - 开发STEM（科学、技术、工程、数学）解题模型 - 训练印度语系语音与音频模型 - 微调视频理解与问答模型 - 开发多语言教育类AI应用 ## 数据采集与版权说明 所有内容均为专有资产，由Prepp内部内容团队与签订雇佣协议的签约学科专家制作。内容合规获取且与课程标准严格匹配，可授权用于商业AI训练场景。 ## 授权与商业获取 本仓库仅包含示例数据。完整的380万条视频数据集可授权用于商业AI训练。 **如需获取授权咨询，请联系：** 安基特·杜贝（Ankit Dubey）—— Collegedunia人工智能数据合作主管 [e-mail] ## 数据集管理方 [Collegedunia Web Private Limited](https://collegedunia.com) | [Prepp](https://prepp.in) 印度哈里亚纳邦古尔冈