yeungchenwa / fontdiffuser Goto Github PK

How long does it take to train this model from scratch?

Hello, I am very interested in your research after reading your paper. But I have a confusion about the second stage of training.

May I ask how to load the model from the first stage during the second stage of training? Is it the total model. pth from the first stage? But it seems that it cannot be loaded. If the downloaded scr_210000.pth can be loaded, what is the relationship between this and the first stage of my own training. Please help answer, thank you very much!!

Thank you very much for your work, I am very interested in this, but I would like to ask whether there are experiments to prove that this model can be applied to handwritten generation?

您好，感谢您优秀的论文。文章中提到三个复杂程度的分类细节见附录，论文结尾并未找到附录，请问可以分享下附录的地址吗？非常感谢！

Thank you for your work, I was very excited when I saw this project on [spaces], and I tried it several times in English and Chinese characters.
The result of the experiment was that the stylistic variation of Chinese characters was very successful, but the English was very poor. I'd like to confirm to you that this is in line with expectations?

• TEST1:
  

• TEST2:
  

when ”Combined with InstructPix2Pix “ releasing ？

您好，这个项目真的太牛了，我想要自己训练一下，但是不知道这个项目数据集的那424种字体的选取有什么特别的考虑吗。还是说从公开字库里随机下载的这些字体并没有什么考虑。如果是有特别进行挑选的话，是个什么挑选策略呢

May I ask whether, during the comparative experiment, all the baselines were trained using the same training set as our method? The article mentions an image size of 96. Regarding DG-Font, whose resolution is limited to 80, how did you address this issue? Did you directly resize it from 80 to 96 for comparison, or did you employ other methods? Thank you.

BTW, without considering the situation of GPU memory, can we set the resolution arbitrarily for our method?

感谢你们的工作！我想用书法数据集来训练一下模型，所以问问模型在毛笔字上的效果如何呢？

在训练的时候默认开启了数据并行导致保存DistributedDataParallel，模型model没有unt这个参数导致出错，如何从并行模型提取需要模型参数

I found that it just generate one character

I am very interested in your repository. The checkpoint file you have provided seems to be trained primarily on Chinese characters, as certain shapes such as the Korean character 'ㅇ' do not render properly.

I was wondering if you happened to have a checkpoint file that has been trained on Korean fonts? Having a model pre-trained on Korean data would be incredibly valuable and helpful for my current project. If such a file is available, I would be most grateful if you could share it with me. If not, I completely understand, and I appreciate you taking the time to consider my request.

Thank you in advance for your help. I look forward to hearing from you.

Hello, how was the SCR_210000 used during the second training process saved? How can I separate the corresponding SCR from the total_model. pth file generated during my training? Looking forward to your answer, thank you very much!!!!