4sp1d3r2/AceCode-V2-122K-lua

该数据集是一个从ACECode-V2-122K衍生而来的转译代码生成数据集。它包含1,800个从Python转换为Lua的编程问题，涵盖：自然语言问题陈述、Lua函数签名、Lua参考解决方案、与luaunit兼容的Lua单元测试，并在Lua 5.4上验证执行。数据集设计用于代码生成研究、语言转译评估、Lua大语言模型基准测试以及Lua能力编码模型的合成训练数据。每个示例源自基于Python的ACECode-V2-122K数据集，并自动转译为惯用的Lua，同时保留问题语义、输入/输出行为和测试正确性。This dataset is a translated code generation dataset derived from the ACECode-V2-122K dataset. It contains 1,800 programming problems translated from Python to Lua, covering natural language problem statements, Lua function signatures, Lua reference solutions, Lua unit tests compatible with luaunit, and validated for execution on Lua 5.4. This dataset is designed for code generation research, language translation evaluation, Lua large language model (LLM) benchmarking, and as synthetic training data for coding models proficient in Lua. Each example originates from the Python-based ACECode-V2-122K dataset and is automatically translated into idiomatic Lua, while retaining the problem semantics, input/output behavior, and test correctness.