[Question] Where to find the pre-trained model and config for Chinese dataset?

[laser33333]

Hi authors,

Thanks for your amazing work and for open-sourcing the code!
I am currently working on my graduation thesis about handwritten text generation and I am very interested in evaluating the Chinese generation capabilities of DiffBrush.

However, in the current repository and the provided Baidu Netdisk/Google Drive links, I could only find the IAM.yml config and a single DiffBrush-ckpt.pt (which seems to be trained on the English dataset). When I input Chinese characters, it returns 0it and skips generation.

Could you please provide or point me to the pre-trained weights and config file (e.g., CASIA) for Chinese text-line generation?

Thank you very much for your time and help!

---

作者您好！非常感谢开源如此优秀的工作！
我目前正在做笔迹鉴定相关的毕业设计，对 DiffBrush 的中文生成效果非常感兴趣。但我发现在目前的仓库和网盘链接中，只看到了针对英文的 IAM.yml 和对应的权重，输入中文时会直接跳过（0it）。
请问在哪里可以获取论文中用于中文数据集（如 CASIA）的配置文件和预训练模型呢？十分感谢您的解答和帮助！

[msyloveldx]

I also urgently need a handwritten Chinese dataset model. Have you found one？