plaguss/test_personahub_fineweb_edu
收藏资源简介：
该数据集是通过distilabel工具创建的，包含一个pipeline.yaml文件，用于复现生成数据集的流程。数据集的每个样本包含text、persona、distilabel_metadata和model_name四个字段。其中，text字段包含一段文本，persona字段描述了文本的生成者角色，distilabel_metadata字段包含生成文本的原始输出，model_name字段指定了生成文本所使用的模型。数据集只有一个配置（default），并且可以通过Hugging Face的datasets库加载。
This dataset was created using the distilabel tool and includes a pipeline.yaml file to reproduce the pipeline that generated it. Each example in the dataset contains four fields: text, persona, distilabel_metadata, and model_name. The text field contains a piece of text, the persona field describes the role of the text generator, the distilabel_metadata field contains the raw output of the generated text, and the model_name field specifies the model used to generate the text. The dataset has only one configuration (default) and can be loaded using the Hugging Face datasets library.
数据集概述
数据集结构
特征
- text: 类型为字符串。
- persona: 类型为字符串。
- distilabel_metadata: 结构化数据，包含以下字段：
- raw_output_text_to_persona_0: 类型为字符串。
- model_name: 类型为字符串。
数据分割
- train: 包含3个样本，占用14689字节。
配置
- default: 包含一个数据文件，路径为
data/train-*。
标签
- synthetic
- distilabel
- rlaif
数据集大小
- 下载大小: 13811字节
- 数据集大小: 14689字节
示例
json { "distilabel_metadata": { "raw_output_text_to_persona_0": "A health-conscious food blogger or a specialty tea shop owner knowledgeable about tea varieties and their benefits." }, "model_name": "meta-llama/Meta-Llama-3-70B-Instruct", "persona": "A health-conscious food blogger or a specialty tea shop owner knowledgeable about tea varieties and their benefits.", "text": "Ever wondered what white tea is? Why we call it white anyway? Here are a few interesting facts about white tea, made from the plant Camellia sinensis: - White tea is similar to green tea in that it has undergone little processing and no fermentation whatsoever. - The leaves are harvested before they open fully, when the unopened buds are still covered by thin silvery white hair. Hence the white tea name. - White tea has a subtle and slightly sweet flavor that sharpens the senses without overwhelming the tongue, whereas green tea has a u201cgrassyu201d taste. - There is 25% to 30% less caffeine in white tea than in green tea, though it depends on the variety. - White tea health benefits are the highest among the tea familyu00b9. - It is more expensive than green or black tea because white tea is scarcer. Yet, at Savorique, white tea is as cheap as black or green tea. Note 1: Science Daily u201cu201dWhite Tea Could Keep You Healthy and Looking Youngu201d", (August 14, 2009) and Science Daily u201cu201dWhite Tea Beats Green Tea In Fighting Germsu201d", (May 28, 2004)." }
加载数据集
python from datasets import load_dataset
ds = load_dataset("plaguss/test_personahub_fineweb_edu")
