jescy525/nexus-sft-v1-sources

这是一个监督微调数据集，属于AETHER家族，属于nexus_core组。数据格式为JSONL ChatML消息，任务类型已标记，并应用了MinHash去重（阈值为0.85）。模式包括消息数组（包含角色如系统、用户或助手及内容）、任务类型（如函数调用、代码、推理链等）、来源数据集、语言（如英语或法语）和系统来源。数据集由prepare_sft.py流水线生成（日期2026-05-25），支持多语言对话，适用于文本生成和对话任务。

This is a supervised fine-tuning dataset from the AETHER family, belonging to the nexus_core group. The format is JSONL ChatML messages, with task_type tagged and MinHash deduplication applied (threshold 0.85). The schema includes messages array (with roles such as system, user, or assistant and content), task_type (e.g., function_calling, code, reasoning_cot), source_ds, lang (e.g., en or fr), and system_source. Generated by the prepare_sft.py pipeline (date 2026-05-25), it supports multilingual conversations and is suitable for text-generation and conversational tasks.