CODEMENV
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/xdshen-ai/Benchmark-of-Code-Migration
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为CODEMENV，包含了922个示例，覆盖了19个Python和Java包。它旨在评估大型语言模型（LLMs）在代码迁移场景下的能力。该数据集的目标是针对代码迁移任务对LLMs进行基准测试，并包含了用于性能评估的指标。规模上，它跨越了19个包，共包含922个示例。任务方面，该数据集包括识别不兼容的函数、检测函数定义的变化，以及适应目标环境的代码调整。
The dataset named CODEMENV contains 922 examples and covers 19 Python and Java packages. It is designed to evaluate the capabilities of Large Language Models (LLMs) in code migration scenarios. The core goal of this dataset is to benchmark LLMs on code migration tasks, and it includes metrics for performance evaluation. In terms of scale, it spans 19 packages and includes a total of 922 examples. For the tasks involved, this dataset covers identifying incompatible functions, detecting changes in function definitions, and making code adjustments to adapt to target environments.
提供机构：
Research team (specific authors not mentioned)
