Weni/Zeroshot_Train-20K_nenhuma_tweet-format

--- configs: - config_name: default data_files: - split: train path: data/train-* dataset_info: features: - name: source_text dtype: string - name: target_text dtype: string splits: - name: train num_bytes: 4411602 num_examples: 20000 download_size: 1748719 dataset_size: 4411602 task_categories: - zero-shot-classification language: - pt size_categories: - 10K<n<100K --- # Dataset Card for "Zeroshot_Train-20K_nenhuma_tweet-format" This dataset is a train dataset for the Zeroshot models. It has 20.000 data in a prompt format exclusively for train with class 'nenhuma' in Brazilian Portuguese. Prompt: ``` "Classifique o tweet entre 'classe1', 'classe2', 'classe3', 'classe4', 'nenhuma' \\n\\nTweet: frase \\n\\nLabel: 'other' ``` The dataset was divided as follows: <br> ``` - 6,000 data: prompt with class option without target class (nenhuma) - 7,000 data: prompt with class option + target class included as an option. target class is not correct - 7,000 data: prompt with class option + target class. target class is correct ``` ## How to load and use this dataset: ``` from datasets import load_dataset dataset = load_dataset("Weni/Zeroshot_Train-20K_nenhuma_tweet-format") dataset ``` [More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)

原始信息汇总

数据集卡片 "Zeroshot_Train-20K_nenhuma_tweet-format"

数据集概述

该数据集是一个用于零样本学习模型的训练数据集，包含20,000条数据，采用巴西葡萄牙语的提示格式，专门用于训练带有nenhuma类别的模型。

数据集结构

特征

• source_text: 字符串类型，源文本。
• target_text: 字符串类型，目标文本。

分割

• train: 训练集，包含4,411,602字节，20,000个样本。

数据集大小

• 下载大小: 1,748,719字节
• 数据集大小: 4,411,602字节

任务类别

• 零样本分类

语言

• 巴西葡萄牙语

数据集大小类别

• 10K<n<100K

数据集划分

• 6,000条数据: 提示包含类别选项，但不包含目标类别（nenhuma）。
• 7,000条数据: 提示包含类别选项和目标类别，但目标类别不正确。
• 7,000条数据: 提示包含类别选项和目标类别，且目标类别正确。

加载和使用

python from datasets import load_dataset dataset = load_dataset("Weni/Zeroshot_Train-20K_nenhuma_tweet-format") dataset