natyu666/SoloAI-SFT-20260501-0136
收藏资源简介:
SoloAI SFT数据集是一个用于SFT微调和指令调优的数据集,包含30条数据,格式为Instruction-Input-Output。数据集支持英文和中文,适用于LLM指令微调、Prompt Engineering研究和多语言支持。每条数据包含任务指令、输入上下文和期望输出三个字段。数据集由SoloAI自动化数据管道生成,从HuggingFace Datasets Hub发现高质量数据集,经过AI清洗和质量过滤后发布。数据集仅供研究和实验用途,商业用途需联系SoloAI进行定制。
The SoloAI SFT Dataset is designed for SFT fine-tuning and instruction tuning, containing 30 data entries in the Instruction-Input-Output format. It supports both English and Chinese and is suitable for LLM instruction tuning, Prompt Engineering research, and multilingual support. Each data entry includes three fields: task instruction, input context, and expected output. The dataset is generated by the SoloAI automated data pipeline, which discovers high-quality datasets from the HuggingFace Datasets Hub, cleans them into the SFT format using AI, and releases them after quality filtering. The dataset is intended for research and experimental purposes only, and commercial use requires customization through SoloAI.




