EgehanEralp/imdb-single-sentence
收藏Hugging Face2024-06-30 更新2024-07-06 收录
下载链接：
https://hf-mirror.com/datasets/EgehanEralp/imdb-single-sentence
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是将原始的IMDb数据集中的多句子评论转换为单句子格式的结果，旨在使基准研究更兼容于包含单句子输入的数据集（如SST-2、HateSpeech、Tweet-Emotion等）。通过微调RoBERTa模型进行情感分析，分割IMDb数据集中的句子，并使用模型选择最具代表性的句子来代表原始评论的情感。
This dataset is derived from the original multi-sentence IMDb dataset (stanfordnlp/imdb) and transformed into a single-sentence format. The transformation process includes fine-tuning with the RoBERTa model, splitting sentences, obtaining sentiment predictions through the model, selecting the sentences with the highest confidence positive or negative labels, and finally creating a single-sentence IMDb dataset where each multi-sentence review is represented by the sentence that best captures the sentiment of the original review.
提供机构：
EgehanEralp
