Dahoas/MATH_train_K_100_qwen_1.5_4B_outputs
收藏Hugging Face2024-10-22 更新2024-12-14 收录
下载链接:
https://hf-mirror.com/datasets/Dahoas/MATH_train_K_100_qwen_1.5_4B_outputs
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含750000个样本,主要用于训练目的。每个样本包含多个特征字段,包括问题(problem)、难度级别(level)、类型(type)、解决方案(solution)、提示(prompt)、响应(response)和推理ID(inference_id)。数据集的总大小为1733627885字节,下载大小为293591320字节。数据集使用默认配置,数据文件路径为data/train-*。
The dataset contains 750,000 samples primarily intended for training purposes. Each sample includes multiple feature fields such as problem, level, type, solution, prompt, response, and inference_id. The total size of the dataset is 1,733,627,885 bytes, with a download size of 293,591,320 bytes. The dataset uses the default configuration, with data files located at data/train-*.
提供机构:
Dahoas



