Wiki-40B
收藏arXiv2025-09-30 收录
下载链接：
https://research.google/pubs/pub49029/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了经过清理的英文维基百科文本，并划分为训练集和验证集。此外，该数据集被用于涉及稳定扩散和OPT语言模型的量化实验。具体任务是对使用英文维基百科子集的量化模型进行评估。
This dataset consists of cleaned English Wikipedia texts, which are split into training and validation sets. Additionally, this dataset has been utilized in quantization experiments involving Stable Diffusion and OPT language models. The specific task is to evaluate quantized models that employ the English Wikipedia subset.
提供机构：
Hugging Face
