Replete-AI/data_processing_code
收藏Hugging Face2024-05-13 更新2024-06-12 收录
下载链接：
https://hf-mirror.com/datasets/Replete-AI/data_processing_code
下载链接
链接失效反馈官方服务：
资源简介：
---
license: apache-2.0
---
These are the scripts I used to clean the Replete-AI/code_bagel and Replete-AI/code_bagel_hermes-2.5 datasets.
In order for these scripts to work your datasets need to be in the format bellow, with the condition that each line is its own .json object.
```
{"instruction": "", "input": "", "output": ""}
{"instruction": "", "input": "", "output": ""}
{"instruction": "", "input": "", "output": ""}
{"instruction": "", "input": "", "output": ""}
{"instruction": "", "input": "", "output": ""}
```
These scripts are used to clean the Replete-AI/code_bagel and Replete-AI/code_bagel_hermes-2.5 datasets. The dataset format requires each line to be an independent JSON object containing instruction, input, and output fields.
提供机构：
Replete-AI
原始信息汇总
数据集概述
数据集用途
- 用于清洗
Repule-AI/code_bagel和Repule-AI/code_bagel_hermes-2.5数据集的脚本。
数据格式要求
- 每行应为一个独立的
.json对象，格式如下： json {"instruction": "", "input": "", "output": ""}
许可证
- 遵循Apache-2.0许可证。
