tanganke/smol-smoltalk-Interaction-SFT-formatted
收藏Hugging Face2025-10-24 更新2025-11-15 收录
下载链接:
https://hf-mirror.com/datasets/tanganke/smol-smoltalk-Interaction-SFT-formatted
下载链接
链接失效反馈官方服务:
资源简介:
ReactiveAI的Smol-Smoltalk交互SFT数据集,用于交互式监督微调反应式转换器模型,特别是RxT-Beta模型。该数据集是基于HuggingFaceTB/smol-smoltalk修改的,去除了系统提示,并将每个用户和助手的消息对作为单独的交互处理。数据集包含训练集和验证集,适用于小规模研究模型。
ReactiveAI Smol-Smoltalk Interaction SFT Dataset, used for Interactive Supervised Fine-Tuning of Reactive Transformer models, especially the RxT-Beta model. This dataset is modified from HuggingFaceTB/smol-smoltalk, excluding system prompts and treating each user-assistant message pair as a separate interaction. The dataset includes training and validation sets, suitable for small-scale research models.
提供机构:
tanganke



