BEE-spoke-data/flan-v2-hf
收藏Hugging Face2025-12-29 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/BEE-spoke-data/flan-v2-hf
下载链接
链接失效反馈官方服务：
资源简介：
flan-v2数据集是一个用于文本生成任务的数据集，包含三个配置：default、en-targets和targets-3w。每个配置的训练集都包含inputs、targets和task三个特征。en-targets配置通过fasttext-langdetect过滤了targets列，仅保留英语且得分大于等于0.65的条目；targets-3w配置则过滤了targets列，仅保留包含3个或更多单词的条目。数据集的总大小和下载大小也被详细列出。
The flan-v2 dataset is designed for text generation tasks and includes three configurations: default, en-targets, and targets-3w. Each configurations training set contains three features: inputs, targets, and task. The en-targets configuration filters the targets column using fasttext-langdetect to retain only English entries with a score of 0.65 or higher. The targets-3w configuration filters the targets column to retain only entries with three or more words. The total size and download size of the dataset are also detailed.
提供机构：
BEE-spoke-data
