DrUkachi/ktt-math-tutor-data

KTT数学辅导数据集是为AIMS KTT黑客马拉松Tier-3提交的数据产物，项目名称为S2.T3.1早期学习者AI数学辅导。数据集包含核心课程和种子数据，其中课程数据有80个项目，涵盖5个子技能(计数、数感、加法、减法和文字问题)，支持英语、法语和Kinyarwanda语言，难度分为1-10级，适合5-9岁不同年龄段。还包括12个手工编写的种子项目、5个快速诊断探针项目、儿童语音样本种子和家长报告模式。此外还包含用于合成儿童语音ASR语料库的训练和评估清单，但实际音频文件需要通过运行代码库中的脚本生成。

The KTT Math Tutor dataset is a data artefact for the AIMS KTT Hackathon Tier-3 submission titled S2.T3.1 AI Math Tutor for Early Learners. It contains core curriculum and seed data, with 80 curriculum items covering 5 sub-skills (counting, number sense, addition, subtraction, word problems) in EN/FR/KIN languages, difficulty levels 1-10, and age bands 5-9. Also includes 12 hand-authored seed items, 5 diagnostic probe items, child utterance samples, and parent report schema. Additionally provides manifests for a synthetic child-voice ASR corpus (training and evaluation), though actual WAV files must be generated by running scripts from the code repository.