jiarui1/Minimind_train_dataset
收藏Hugging Face2025-03-04 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/jiarui1/Minimind_train_dataset
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了多个文件，分别是：中英文混合数据集，由五个文件随机组合而成，适用于多语言模型微调；Llama-3.3-70B、Llama-3.1-70B、Llama-3.0-8B-Instruct和Llama-3.0-70B数据集，来源于Magpie-Llama的不同版本，经过筛选条件处理；灵智数据集，来源于Lingzhi-SFT-data-61w；预训练混合数据集，由匠数数据集中的中文英文数据集选取并拼接而成。
The dataset consists of multiple files: a mixed Chinese-English dataset composed of five files for multilingual model fine-tuning; Llama-3.3-70B, Llama-3.1-70B, Llama-3.0-8B-Instruct, and Llama-3.0-70B datasets from different versions of Magpie-Llama, processed by filtering conditions; the Lingzhi dataset from Lingzhi-SFT-data-61w; and a pre-training mixed dataset composed of selected and concatenated Chinese and English datasets from the Craftsmanship dataset.
提供机构：
jiarui1
