fewshot-font-generation
fewshot-font-generation copied to clipboard
clovaai
The unified repository for few-shot font generation methods. This repository includes FUNIT (ICCV'19), DM-Font (ECCV'20), LF-Font (AAAI'21) and MX-Font (ICCV'21).
Use the MX-Font. The source image is black on white. But result/mx/images/last.png is like this: 
What are unseen_chars and seen_chars? What is the difference between source_path and data_dir? thanks
When using the FUNIT model, my original style font image i   s There are two kinds of target style images, respectively   ser-images.githubusercontent.com/60809733/169986605-226ea8f5-ba73-47e6-bd44-42377ef7a09e.jpg)  But why do...
Hello, may I ask why the result of image/result in the training process is good, but the test result is poor, and the images used are the same? The first...
Hello, I trained 98 korean fonts more than 100,000 times. If I try to convert my own handwriting into letters, the recognition rate drops considerably. If there is a great...
Hello, I used the mx-font training data set in this project. There are 36 fonts in total, including 13 XIAOZHUANTI, and 800 pictures for each font. However, during the 14-15w...
Thank you very much for your excellent work. I've spent some time studying and learning from it. However, I've encountered a problem that I'd like to ask about. While testing...
Thanks for the great work, I have the following questions 1. I have trained 30 fonts, some images of the training process are as follows, I don’t think 65w iter...
Hi, I can't find the weights of the classifiers trained with FTransGAN Dataset. (weights/evaluator_ftransgan_20epoch.pth), can you provide the weights？ Thanks a lot.
I use LF-Font for training and have the following problem: 1. Are all the unicode encodings in the given example data/chn/train_chars.json used for training? I think the number of characters...