WeiChow/cc3m_hed

CC3M是一个包含图像数据的英文数据集，适用于文本到图像和图像到图像的任务。数据集包含id和image两个字段，其中image字段包含图片数据。整个数据集大小约为173.9GB，训练集包含约301.7万个示例。数据集的图片经过了处理，确保宽度或高度不会超过1024像素。

CC3M is an English dataset containing image data, suitable for text-to-image and image-to-image tasks. The dataset includes id and image fields, with the image field containing image data. The entire dataset is approximately 173.9GB in size, with the training set containing about 3,016,640 examples. The images in the dataset have been processed to ensure that neither the width nor the height exceeds 1024 pixels.