Django

该数据集包含了18,805行Python源代码，每行代码都配有一个自然语言表述。此外，该数据集被划分为训练集、验证集和测试集，分别包含16,000、1,000和1,805个实例。规模上，总共有18,805行代码，任务类型是代码生成。

This dataset contains 18,805 lines of Python source code, with each line paired with a corresponding natural language description. Furthermore, the dataset is split into training, validation, and test sets, which contain 16,000, 1,000, and 1,805 instances respectively. In terms of scale, the total number of code lines is 18,805, and the task type of this dataset is code generation.