AbstractPhil/human-templated-captions-1b
收藏Hugging Face2025-05-20 更新2025-08-30 收录
下载链接：
https://hf-mirror.com/datasets/AbstractPhil/human-templated-captions-1b
下载链接
链接失效反馈官方服务：
资源简介：
这是一个适用于文本生成和文本到文本生成任务的英语数据集，数据量在100M到1B之间。数据集使用了MIT许可证，当前存在CSV文件读取问题，未来将提供Parquet格式的文件分割，以及为延伸训练过程设计的长标题分割。
This is an English dataset for text generation and text-to-text generation tasks, with a size ranging from 100M to 1B. The dataset is licensed under MIT. There is currently an issue with reading CSV files, and in the future, file splits in Parquet format will be provided, as well as long caption splits designed for the elongation training process.
提供机构：
AbstractPhil
