LogicStar/SWE-Star
收藏Hugging Face2026-01-26 更新2026-02-07 收录
下载链接：
https://hf-mirror.com/datasets/LogicStar/SWE-Star
下载链接
链接失效反馈官方服务：
资源简介：
SWE-Star是一个基于Qwen2.5-Coder系列的语言模型训练数据集，包含约25万个从Devstral-2-Small通过SWE-Smith任务提炼的代理编码轨迹。该数据集用于训练和评估语言模型，特别是在Python代理编码方面的性能。数据集生成、训练和评估的完整流程在GitHub仓库中公开，可供复现结果。
SWE-Star is a family of language models based on the Qwen2.5-Coder family and trained on the SWE-Star dataset, which contains approximately 250k agentic coding trajectories distilled from Devstral-2-Small using SWE-Smith tasks. The dataset is designed for training and evaluating language models, particularly in the context of agentic coding in Python. The complete data generation, training, and evaluation pipeline is openly available in the GitHub repository for reproducibility.
提供机构：
LogicStar
