jescy525/aether-sft-v2-code

AETHER家族SFT数据集——代码组。这是一个用于监督微调（SFT）的数据集，专注于代码相关任务。数据格式为JSONL ChatML消息，包含任务类型标记（如函数调用、代码、推理链等），并应用了MinHash去重（阈值为0.85）。数据集支持英语和法语，来源于其他HuggingFace数据集，系统来源包括默认和覆盖设置。数据规模在10万到100万之间，适用于文本生成和对话任务。AETHER Family SFT Dataset - Code Subset. This is a supervised fine-tuning (SFT) dataset focused on code-related tasks. The dataset adopts the JSONL ChatML message format, which includes task type tags such as function calls, code, reasoning chains, etc., and has undergone MinHash deduplication with a threshold of 0.85. It supports both English and French, is derived from other Hugging Face datasets, and its system sources cover both default and override settings. The dataset size ranges from 100,000 to 1,000,000 samples, and is applicable to text generation and conversational tasks.