stefanocarrera/autophagycode_D_he_train-mercury_Qwen3-4B_strategy_trust_t1.1_g9_run0_metrics
收藏资源简介:
该数据集包含164个训练示例,总大小为234931字节,下载大小为102215字节。数据集的特征涵盖了任务标识、入口点、可执行性、正确性、测试通过/失败数量、测试运行时间(当前为null值)、错误类型、多种代码复杂度度量(如Halstead度量、圈复杂度、可维护性指数)、代码行数(包括总行数和源代码行数)、注释百分比、TTR(类型标记比率)、令牌字典、香农熵、预测熵(均值和最大值)、定义函数数量以及入口点重复性等字段。这些特征主要用于代码分析和评估任务,可能涉及代码质量、执行结果或机器学习模型预测。
This dataset comprises 164 training instances, with a total size of 234,931 bytes and a download size of 102,215 bytes. The features of this dataset encompass fields including task identifier, entry point, executability, correctness, number of passed/failed tests, test runtime (currently null), error type, various code complexity metrics (e.g., Halstead metrics, cyclomatic complexity, maintainability index), lines of code (including total lines and source code lines), comment percentage, TTR (Type-Token Ratio), token dictionary, Shannon entropy, predictive entropy (mean and maximum values), number of defined functions, and entry point repetition. These features are predominantly utilized for code analysis and evaluation tasks, which may involve code quality, execution results, or machine learning model prediction.




