DS-1000

该数据集是一个包含了1000个独立数据科学工作流的基准测试集，这些工作流跨越了七个不同的库，旨在评估代码生成在数据科学工作流中的性能表现。规模达到了1000个工作流，任务专注于数据科学工作流中的代码生成。This dataset is a benchmark test set comprising 1000 independent data science workflows spanning seven distinct libraries, designed to evaluate the performance of code generation in data science workflows. With a total of 1000 workflows, the tasks of this dataset focus on code generation within data science workflows.