nvidia/OpenCodeGeneticInstruct
收藏Hugging Face2025-05-23 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/nvidia/OpenCodeGeneticInstruct
下载链接
链接失效反馈官方服务：
资源简介：
OpenCodeGeneticInstruct是一个大规模的数据集，包含超过1500万个用遗传指令方法合成的Python编码指令。该数据集用于监督微调大型语言模型，以提高其代码生成能力。每个样本包括一个编码问题/指令及其相应的答案，答案包含Python语言的编码解决方案。
OpenCodeGeneticInstruct is a large-scale dataset containing more than 15 million synthetic Python coding instructions generated using the Genetic-Instruct approach. This dataset is used for supervised fine-tuning of large language models (LLMs) to enhance their code generation capabilities. Each sample includes a coding question/instruction and its corresponding answer, which is the coding solution in Python.
提供机构：
nvidia
