jschoormans/sd3_v6
收藏Hugging Face2025-02-14 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/jschoormans/sd3_v6
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了三种配置的数据，每种配置下都包含图像、条件图像、深度图像、条件深度图像、文本、空文本、图像质量评价指标和视场信息。图像和条件图像的数据类型为image，深度图像的数据类型为1024x1024的float16数组，文本和空文本的数据类型为字符串，图像质量评价指标的数据类型为float32，视场信息的数据类型为int32。每种配置的训练集包含不同数量的示例，数据集大小和下载大小也有所不同。
The dataset consists of three configurations, each containing images, conditional images, depth images, conditional depth images, text, empty text, image quality evaluation metrics, and field of view information. The data types for images and conditional images are image, for depth images are 1024x1024 float16 arrays, for text and empty text are strings, for image quality evaluation metrics are float32, and for field of view information are int32. Each configurations training set contains a different number of examples, and the dataset size and download size also vary.
提供机构：
jschoormans
