Description
Hello, author and fellow experts. I would greatly appreciate your assistance!
I have some basic operational questions regarding StyleGAN3, and I hope you can provide some guidance.
I aim to generate new faces using the ffhq-256*256 dataset. Could you please advise me on the steps to follow? (Specifically, should I first train the dataset using train.py, and then use the generated weight file .pkl in the gen-images.py file to create new faces?)
How can I convert the quality metrics, such as fid50kfull and kid50kfull, generated after running train.py, into a chart format for a more intuitive observation of the results?
Thank you very much for your patience and assistance!
作者您好，我想寻求您的帮助！
我想咨询一些关于StyleGAN3的基本操作问题，请您多包涵指教。
1、我想用ffhq-256*256数据集最终生成一些新的人脸，我应该怎么样操作？（是先使用 train.py 训练数据集，然后在用生成的权重文件 .pkl 放入gen-images.py文件中生成新的人脸吗）
2、运行 train.py 后产生的fid50k_full、kid50k_full等质量指标，怎么使其变成图表的形式，直观的观察结果？
非常感谢您耐心的解答！！！