khanhromvn/Atri-QA-Dataset-Vi-100K

Atri-QA-Dataset-Vi-100K 是一个高质量的越南语对话数据集，包含 100,000 个问题-回答对，涵盖了日常对话、情感支持、讲故事等多个主题。该数据集专为训练具有一致、同情和友好性格的对话式 AI 模型而设计，灵感来源于动画角色 Atri。数据由 GPT-4 生成，并经过严格的自动过滤和人工审核的质量控制流程。该数据集适用于微调对话模型、指令微调、聊天机器人训练和越南语自然语言处理研究。

Atri-QA-Dataset-Vi-100K is a high-quality Vietnamese conversational dataset consisting of 100,000 question-answer pairs across various themes such as casual conversation, emotional support, storytelling, and more. Designed to train conversational AI models with a consistent, empathetic, and friendly personality inspired by the anime character Atri, the data is generated using GPT-4 and undergoes a rigorous quality control process involving automated filters and human review. The dataset is suitable for fine-tuning conversational models, instruction tuning, chatbot training, and Vietnamese NLP research.