mlabonne/dolphin-r1-deepseek
收藏Hugging Face2025-01-30 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/mlabonne/dolphin-r1-deepseek
下载链接
链接失效反馈官方服务：
资源简介：
Dolphin R1 DeepSeek是一个用于训练R1风格推理模型的Apache-2.0许可数据集。该数据集是DeepSeek子集的重新格式化版本，它将模型在对话中的响应添加了特殊标记，包括<|begin_of_thought|>，<|end_of_thought|>，<|begin_of_solution|>，<|end_of_solution|>，以方便使用。数据集包含训练集，其中有300,018个示例，总大小为3,776,761,407字节。
Dolphin R1 DeepSeek is an Apache-2.0 licensed dataset for training R1-style reasoning models. This dataset is a reformatted version of the DeepSeek subset, which adds special tokens for the models response in the conversation, including <|begin_of_thought|>, <|end_of_thought|>, <|begin_of_solution|>, <|end_of_solution|>, for ease of use. The dataset includes a training set with 300,018 examples, totaling 3,776,761,407 bytes.
提供机构：
mlabonne
