alwaysgood/Bloomberg_Financial_News_processed
收藏资源简介:
该数据集是经过处理的彭博金融新闻数据集,原始数据来源于`danidanou/Bloomberg_Financial_News`。数据集经过多种文本处理步骤,包括流式处理、文本字段提取、换行符规范化、控制字符移除、空格转换、轻量级乱码恢复、空白规范化等。处理后的数据集包含`text`、`language`、`style_tag`和`metadata`字段,并以`jsonl.gz`格式输出。数据集主要用于文本生成任务,包含446,707条处理后的新闻条目。
This dataset is a processed version of Bloomberg Financial News, originally sourced from `danidanou/Bloomberg_Financial_News`. The dataset has undergone various text processing steps, including streaming, text field extraction, newline normalization, control character removal, space conversion, lightweight mojibake recovery, and whitespace normalization. The processed dataset retains fields such as `text`, `language`, `style_tag`, and `metadata`, and is output in `jsonl.gz` format. The dataset is primarily used for text generation tasks and contains 446,707 processed news entries.




