PJMixers-Dev/anthracite-org_c2_logs_32k_llama3_qwen2_v1.3-qwq-all-aphrodite

这是一个用于训练对话模型的32k条日志数据集，基于llama3和qwen2，版本为v1.3。数据集被设计为只使用最后一轮对话内容进行训练，以教会模型始终输出思考内容。数据集的格式与R1模型的多轮对话训练格式相兼容。

This is a 32k log dataset for training dialogue models, based on llama3 and qwen2, version v1.3. The dataset is designed to train using only the last round of conversation content, to teach the model to always output thinking content. The format of the dataset is compatible with the multi-round conversation training format of the R1 model.