NancyT/tds-sft
收藏官方服务：
资源简介：
该数据集是一个包含指令-输入-响应结构的大规模训练数据集，用于训练对话或指令跟随模型。数据特征包括instruction（指令）、input（输入）和response（响应）三个文本字段。数据集仅包含训练分割，包含约1101万条样本，总大小约3.14GB，下载大小约2.1GB。
This dataset is a large-scale training dataset with instruction-input-response structure, designed for training dialogue or instruction-following models. The features include three text fields: instruction, input, and response. The dataset contains only the training split, with approximately 11.02 million examples, total size of about 3.14GB, and download size of about 2.1GB.
提供机构：
NancyT
