Ling-Coder-SFT

This dataset is part of the SFT data used for fine-tuning the Ling-Coder Lite model. It includes over 5 million samples covering more than 20 programming languages and spans various topics such as text-to-code, code execution reasoning, and code completion.

本数据集为用于微调Ling-Coder Lite模型的监督微调（Supervised Fine-Tuning，SFT）数据的一部分。该数据集包含超500万条样本，覆盖20余种编程语言，涵盖文本转代码（text-to-code）、代码执行推理（code execution reasoning）以及代码补全（code completion）等多类任务主题。