LLMTeamAkiyama/cleaned_nvidia_OpenCodeReasoning
收藏Hugging Face2025-08-05 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/LLMTeamAkiyama/cleaned_nvidia_OpenCodeReasoning
下载链接
链接失效反馈官方服务：
资源简介：
OpenCodeReasoning数据集是一个用于代码推理任务的数据集，包含11,275个样本，以JSONL格式存储，总token数为126,859,041。该数据集经过去重和去除极端值等预处理，适用于竞赛类别，难度评分为15，遵循MIT和CC-BY-4.0双许可证。
The OpenCodeReasoning dataset is a dataset for code reasoning tasks, containing 11,275 samples stored in JSONL format, with a total of 126,859,041 tokens. The dataset has been preprocessed to remove duplicates and extreme values, is suitable for competition categories, has a difficulty score of 15, and is licensed under both MIT and CC-BY-4.0.
提供机构：
LLMTeamAkiyama
