diffusion-cot/gpt-edit-simpler

这是一个包含图像输入和输出的数据集，旨在进行某种编辑任务。数据集中的每个示例都有一个唯一的标识符(id)，并且来源于不同的编辑器(hqedit、omniedit、ultraedit)。此外，每个示例还包含了任务类型(task)、输入图像的描述(input_description)、原始指令(instruction_original)、以及指令类型(instruction_type，分为original、rewrite、complex三种)。数据集被拆分为训练集(train)，包含约1553575个示例，总大小约为4410.23GB。

This is a dataset containing image inputs and outputs for some editing task. Each example in the dataset has a unique identifier (id) and originates from different editors (hqedit, omniedit, ultraedit). Additionally, each example includes the task type (task), a description of the input image (input_description), the original instruction (instruction_original), and the type of instruction (instruction_type, divided into original, rewrite, and complex). The dataset is split into a training set (train) containing approximately 1,553,575 examples, with a total size of about 4,410.23GB.