StarCoder 2 与 The Stack v2:下一代
StarCoder 2 and The Stack v2: The Next Generation
AI总结:
本文介绍 BigCode 的 StarCoder2 系列与 The Stack v2 数据集,基于多源高质量代码数据训练 3B、7B、15B 模型,并证明其在多项代码、数学和推理基准上的优势及开放透明性。
摘要:
BigCode 项目是一个开放科学合作项目,专注于负责任地开发代码大型语言模型(Code LLM),现推出 StarCoder2。我们与 Software Heritage(SWH)合作,在其源代码档案这一数字公地的基础上构建 The Stack v2。除了涵盖 619 种编程语言的 SWH 代码仓库外,我们还精心挑选了其他高质量数据源,例如 GitHub pull request、Kaggle notebook 和代码文档。由此得到的训练集比第一代 StarCoder 数据集大 4 倍。我们在 3.3 至 4.3 万亿个 token 上训练了具有 3B、7B 和 15B 参数的 StarCoder2 模型,并在一套全面的 Code LLM 基准测试上对其进行了充分评估。我们发现,小型模型 StarCoder2-3B 在大多数基准测试上优于其他类似规模的 Code LLM,并且也优于 StarCoderBase-15B。大型模型 StarCoder2-15B 显著优于其他可比规模的模型。此外,它达到或超过 CodeLlama-34B,而后者规模是其两倍多。尽管 DeepSeekCoder-33B 在高资源语言的代码补全任务上表现最佳,但我们发现 StarCoder2-15B 在数学和代码推理基准测试以及若干低资源语言上优于它。我们以 OpenRAIL 许可证开放模型权重,并通过发布源代码数据的 SoftWare Heritage persistent IDentifier(SWHID)来确保训练数据的完全透明。
英文摘要:
The BigCode project, an open-scientific collaboration focused on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder2. In partnership with Software Heritage (SWH), we build The Stack v2 on top of the digital commons of their source code archive. Alongside the SWH repositories spanning 619 programming languages, we carefully select other high-quality data sources, such as GitHub pull requests, Kaggle notebooks, and code documentation. This results in a training set that is 4x larger than the first StarCoder dataset. We train StarCoder2 models with 3B, 7B, and 15B parameters on 3.3 to 4.3 trillion tokens and thoroughly evaluate them on a comprehensive set of Code LLM benchmarks. We find that our small model, StarCoder2-3B, outperforms other Code LLMs of similar size on most benchmarks, and also outperforms StarCoderBase-15B. Our large model, StarCoder2- 15B, significantly outperforms other models of comparable size. In addition, it matches or outperforms CodeLlama-34B, a model more than twice its size. Although DeepSeekCoder- 33B is the best-performing model at code completion for high-resource languages, we find that StarCoder2-15B outperforms it on math and code reasoning benchmarks, as well as several low-resource languages. We make the model weights available under an OpenRAIL license and ensure full transparency regarding the training data by releasing the SoftWare Heritage persistent IDentifiers (SWHIDs) of the source code data.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。