usamakenway/unified-uncensored-qwen-chatml-sft

该数据集是一个混合许可证的指令/聊天数据集编译，已转换为单一的Qwen/ChatML风格的text JSONL格式。每行数据包含text字段（以特定标记格式表示用户和助手的对话）、source（数据集来源）、source_format（原始格式如alpaca、sharegpt等）、source_license（许可证类型如apache-2.0、mit等）和source_family（来源族系）。这是一个混合许可证编译，每行保留原始来源和许可证元数据，用户需负责遵守每个源数据集的条款。整个数据集不一定可商业使用，商业用途需创建过滤导出，仅包含明确许可的源数据并审查归属要求。数据处理包括归一化为中间messages格式、渲染为Qwen/ChatMLtext，并进行精确和近重复检查的去重。This dataset is a mixed-license compilation of instruction/chat datasets converted into a single Qwen/ChatML-style text JSONL format. Each row has a text field with user-assistant dialogue in specific markers, along with metadata for source, source format (e.g., alpaca, sharegpt), source license (e.g., apache-2.0, mit), and source family. It is a mixed-license compilation where each row preserves original source and license metadata, and users are responsible for complying with the terms of each source dataset. The full dataset may not be commercially usable; for commercial use, create a filtered export with only permissively licensed sources and review attribution requirements. Processing involves normalization to intermediate messages, rendering to Qwen/ChatML text, and deduplication with exact and near-duplicate checks.