nraptisss/TMF921-intent-to-config-25k

TMF921 Intent-to-Config Dataset (25K) 是一个用于训练大型语言模型（LLMs）将自然语言网络意图转换为符合规范的5G/6G配置的开放数据集。数据集包含25,000个样本（22,500个训练样本和2,500个测试样本），覆盖了6个目标规范层（如TMF921、3GPP TS 28.312、CAMARA等）和8个生命周期操作。所有配置都基于真实的电信标准，如3GPP、TM Forum、ETSI、O-RAN和CAMARA。数据集的特点是全面、开放，并且是首个基于TMF921 v4 API的数据集，填补了现有数据集的空白。

The TMF921 Intent-to-Config Dataset (25K) is an open dataset designed for training large language models (LLMs) to translate natural language network intents into spec-compliant 5G/6G configurations. It includes 25,000 samples (22,500 for training and 2,500 for testing) across 6 target specification layers (e.g., TMF921, 3GPP TS 28.312, CAMARA) and 8 lifecycle operations. All configurations are grounded in real telecom standards such as 3GPP, TM Forum, ETSI, O-RAN, and CAMARA. The dataset is comprehensive, open, and the first to be based on the TMF921 v4 API, filling a gap in existing datasets.