Etherll/code-FIM
收藏Hugging Face2024-07-20 更新2024-07-22 收录
下载链接：
https://hf-mirror.com/datasets/Etherll/code-FIM
下载链接
链接失效反馈官方服务：
资源简介：
该数据集旨在解决开源社区中缺乏良好代码填充数据集和模型的问题。数据集包含了来自多种编程语言的10k代码文件，具体分布为Python 3000个、JavaScript 1500个、TypeScript 1250个、Rust 1750个、Java 1500个、GO 500个和C++ 500个。数据集支持最多4行的代码补全，并提供了训练建议和Colab笔记本链接。
The dataset includes two main features: prompt and completion, both of which are large string types. The dataset is divided into two parts: train and variation, containing 4,472,370 and 10,000 samples respectively. The dataset consists of code files from multiple programming languages, including Python, JavaScript, TypeScript, Rust, Java, GO, and C++. It is primarily used for code completion, supporting up to 4 lines of code completion tasks.
提供机构：
Etherll
