joshuasundance/python-code-instructions-85k-mypo-qaqc

这是一个名为joshuasundance/python-code-instructions-85k-mypo QA/QC artifact的Python代码指令数据集，包含85k条数据。它是使用myponline工具生成的QA/QC衍生品，包含原始数据及经过不同级别过滤的数据：1) 通过基本结构QA/QC检查的filtered_basic数据；2) 通过更严格检查(包括ruff和mypy --strict)的filtered_strict数据。数据集包含训练集(45100条)、验证集(922条)和测试集(438条)，并提供了详细的来源信息和分割统计。需要注意的是，根数据集是经过注释而非清理的，下游用户应根据自己的使用场景选择合适的QA/QC过滤标志。

This is a dataset named joshuasundance/python-code-instructions-85k-mypo QA/QC artifact, containing 85k Python code instructions. It is a QA/QC derivative generated by the `myponline` tool, including original data and data filtered at different levels: 1) `filtered_basic` with rows that pass structural QA/QC checks; 2) `filtered_strict` with rows that pass stricter checks including `ruff` and `mypy --strict`. The dataset contains train (45,100 rows), validation (922 rows), and test (438 rows) splits, with detailed provenance information and split statistics. Note that the root dataset is annotated, not cleaned, and downstream consumers should filter using explicit QA/QC flags that fit their use case.