homebrewltd/Ichigo-pretrain-tokenized-v0.1

本数据集包含四个子数据集：instruction_speech_v1、libris_r_filtered、mls_eng_10k和vivoice。每个子数据集都包括以下特征：文本记录（transcript）、提示语（prompt）、压缩提示语（compress_prompt）和对话（conversations），对话中包含对话内容和角色信息。每个子数据集都有训练集的分割信息，包括字节数和示例数量。此外，还提供了每个数据集的下载大小和总大小。

The dataset consists of four sub-datasets: instruction_speech_v1, libris_r_filtered, mls_eng_10k, and vivoice. Each sub-dataset includes features such as transcript, prompt, compressed prompt, and conversations which contain content and role fields. There are split details for each sub-dataset indicating the size of the training set in bytes and the number of examples. Moreover, the download size and the total size of each dataset are provided.