dailenson / SDT

This repository is the official implementation of Disentangling Writer and Character Styles for Handwriting Generation (CVPR23).
MIT License
932 stars 77 forks source link

关于预训练模型的训练方法的一个疑问 #81

Open mmybit opened 1 month ago

mmybit commented 1 month ago

作者您好,我想询问您一下发布的预训练模型是只在 CASIA-OLHWDB (1.0-1.2)上训练吗,论文里写的在测试集 ICDAR-2013 抽80%训练是什么意思呀,是另一个实验设置吗?谢谢您! To demonstrate the superiority of the proposed offline-to-offline handwriting generation framework, we use the offline character images of ICDAR-2013 competition database [54], which contains 60 writers and 3755 different Chinese characters for each writer. We randomly select 80% of the entire dataset as the training set, and the remaining 20% as the test set.

dailenson commented 1 month ago

是的,您贴这段实验描述不是我们的主实验,是另一个拓展实验。我们的主实验做online手写生成的,我们也将其拓展到offline手写生成,您贴的这段就是offline手写生成的实验描述。主实验的描述请看正文的实验细节那里。