SandLogicTechnologies/Indic_Chat_Dataset
收藏Hugging Face2025-09-11 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/SandLogicTechnologies/Indic_Chat_Dataset
下载链接
链接失效反馈官方服务：
资源简介：
IndicMoE多语言数据集是一个大规模的语料库，覆盖了13种印度语言，专门为微调大型语言模型（LLM）而设计。该数据集从53个开源数据集中筛选并整理出约29百万高质量样本，用于自然语言理解（NLU）和自然语言生成（NLG）任务的性能提升，支持指令调整、对话建模、翻译和多语言微调等应用。
The IndicMoE Multilingual Dataset is a large-scale corpus spanning 13 Indian languages, curated specifically for finetuning large language models (LLMs) on Indic languages. The dataset consists of approximately 29 million high-quality samples after filtering and curation from 53 open-source datasets, designed to enhance the performance of LLMs in natural language understanding (NLU) and natural language generation (NLG) tasks, supporting applications such as instruction tuning, dialogue modeling, translation, and multilingual finetuning.
提供机构：
SandLogicTechnologies
