d0rj/ru-instruct
收藏Hugging Face2024-07-08 更新2024-07-22 收录
下载链接:
https://hf-mirror.com/datasets/d0rj/ru-instruct
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是由多个流行的数据集自动翻译并组合而成,经过过滤和去重处理。数据集主要用于文本生成和文本到文本生成任务,包含俄语和代码语言。数据集的特征包括对话内容和角色,以及数据来源。数据集的训练集包含753,650个样本,总大小为1,874,771,723字节。
The dataset is a combination of several popular datasets that have been automatically translated into Russian. It includes datasets like OpenOrca-ru, OpenHermes-2.5-ru, dolphin-ru, alpaca-cleaned-ru, gsm8k-ru, boolq-ru, and conala-mined-ru. The dataset has been filtered for translation artifacts using the Den4ikAI/nonsense_gibberish_detector model and deduplicated using SimHash. It is primarily used for text generation and text-to-text generation tasks. The dataset contains features such as conversations (with roles and content) and source information. It has a training split with 753,650 examples and is tagged as instruct, synthetic, sft, and chat.
提供机构:
d0rj
原始信息汇总
数据集概述
基本信息
- 名称: Russian Instruct Translated
- 语言: 俄语 (ru), 代码 (code)
- 许可: Apache 2.0
- 多语言性: 单语种 (monolingual)
- 数据规模: 100K < n < 1M
数据来源
- 原始数据集:
- d0rj/OpenOrca-ru
- d0rj/OpenHermes-2.5-ru
- d0rj/dolphin-ru
- d0rj/alpaca-cleaned-ru
- d0rj/gsm8k-ru
- d0rj/boolq-ru
- d0rj/conala-mined-ru
任务类别
- 文本生成 (text-generation)
- 文本到文本生成 (text2text-generation)
数据集结构
- 特征:
- conversations:
- content (string)
- role (string)
- source (string)
- conversations:
数据分割
- 训练集:
- 样本数: 753650
- 字节数: 1874771723.0
下载信息
- 下载大小: 839357336
- 数据集大小: 1874771723.0
配置
- 默认配置:
- 数据文件路径: data/train-*
标签
- instruct
- synthetic
- sft
- chat
注意事项
- 存在未解决的翻译问题,主要涉及类似“如何用阿拉伯语说N”的指令。



