KavinduHansaka/Python_Refactor_Dataset
收藏Hugging Face2025-10-09 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/KavinduHansaka/Python_Refactor_Dataset
下载链接
链接失效反馈官方服务：
资源简介：
Python Refactor Dataset是一个包含45,000个合成Python代码重构示例的数据集，用于指导指令微调代码模型。该数据集包含了改进代码可读性、可维护性和风格的示例，同时保持原有功能不变。每个示例都包含了唯一的UUID、自然语言重构指令、重构约束、原始代码片段、目标重构代码、重构理由、语言标签、重构类别标签以及为IBM Granite和Meta CodeLlama模型定制的提示格式。
The Python Refactor Dataset is a collection of 45,000 synthetic Python code refactoring examples designed for guiding instruction-tuning code models. The dataset includes examples that improve code readability, maintainability, and style while preserving the original functionality. Each example contains a unique UUID, natural-language refactoring instructions, refactoring constraints, original code snippets, target refactored code, reasons for changes, language tags, refactoring category tags, and prompt formats tailored for the IBM Granite and Meta CodeLlama models.
提供机构：
KavinduHansaka
