ewrfcas

https://github.com/nlpdata/c3 可以到这里下载对应测试集

是的

Of course you can train it in FFHQ dataset https://github.com/NVlabs/ffhq-dataset

See gen_lr_imgs.py.

Which line does it happens? However, if you use the input size different from the default size, you need to retrain the model.

别急，先看看有没有按照readme操作，然后看看运行是否有什么error或者warning

It seems like 'img' is None

Can you provide more details?

Thanks!

Thanks