apipack/API-Pack-Dataset

API Pack是一个大规模的多编程语言数据集，包含超过100万个指令-API调用对，旨在提高大型语言模型的API调用生成能力。该数据集中的每个实例包括多个字段，如api_name、api_description、api_call_data（进一步包括api_call、lang、functionality、api_arguments、description、domain和path）、instruction、instruction_test、input、output和unique_id。该数据集支持多种编程语言，包括curl、go、java、javascript、libcurl、php、python、ruby和swift。如果发现该数据集有用，请考虑引用相关论文。

API Pack is a massive multi-programming language dataset containing more than 1 million instruction-API call pairs to improve the API call generation capabilities of large language models. Each instance in the dataset includes various fields such as api_name, api_description, api_call_data (which further includes api_call, lang, functionality, api_arguments, description, domain, and path), instruction, instruction_test, input, output, and unique_id. The dataset supports multiple programming languages including curl, go, java, javascript, libcurl, php, python, ruby, and swift. If the dataset is found useful, please consider citing the related paper.

原始信息汇总

数据集详情

许可证信息

• 许可证类型：CC-BY-4.0