PashtoNanoChat
收藏资源简介:
PashtoNanoChat 是一个高质量的普什图语对话数据集,专为聊天机器人训练、对话建模和普什图语大语言模型对齐而设计。该数据集采用OpenAI/Ministral消息格式,包含结构化的普什图语聊天对。每个数据条目以JSON格式存储,包含messages字段,该字段是一个由多个对话轮次组成的列表,每个轮次包含role(取值user或assistant)和content(普什图语文本)两个字段。数据集适用于多种任务,包括普什图语聊天机器人开发、指令调优的大语言模型训练、对话代理构建、文化对齐研究以及低资源自然语言处理研究。数据集以Apache License 2.0许可证发布,可通过Hugging Face的datasets库直接加载使用。
PashtoNanoChat is a high-quality Pashto dialogue dataset designed for chatbot training, dialogue modeling, and alignment of large language models for Pashto. It adopts the OpenAI/Ministral message format and contains structured Pashto chat pairs. Each data entry is stored in JSON format with a messages field, which is a list of multiple dialogue turns; each turn includes role (either user or assistant) and content (Pashto text) fields. The dataset is suitable for various tasks, including Pashto chatbot development, instruction tuning for large language models, dialogue agent construction, cultural alignment research, and low-resource natural language processing studies. It is released under the Apache License 2.0 and can be directly loaded via the Hugging Face datasets library.




