jan-hq/instruction-convert-audio-whispervq-llama3.2-compress
收藏Hugging Face2024-12-09 更新2024-12-14 收录
下载链接:
https://hf-mirror.com/datasets/jan-hq/instruction-convert-audio-whispervq-llama3.2-compress
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含多个特征,包括transcript(转录文本)、text(文本)、compress_tokens(压缩标记)和conversations(对话)。conversations是一个列表,包含content(内容)和role(角色)两个字段。数据集仅包含一个训练集(train),共有3,707,675个样本,总大小为49,099,019,908字节,下载大小为11,985,711,300字节。数据集的配置文件指定了默认配置,数据文件路径为data/train-*。
The dataset contains multiple features, including transcript, text, compress_tokens, and conversations. Conversations is a list containing two fields: content and role. The dataset includes only a training set (train) with 3,707,675 samples, totaling 49,099,019,908 bytes in size and 11,985,711,300 bytes in download size. The datasets configuration file specifies the default configuration, with data file paths as data/train-*.
提供机构:
jan-hq



