how to evaluate performance quantitatively?

[harryxu-yscz]

Thank you for sharing your work! The code runs smooth and I am able to reproduce some good results.

My question is that how to tell from the losses to decide that the model is well-trained? I have read your paper but it is still not quite clear to me. I am new to GAN and I hope there is some metric like "accuracy" for classification tasks that can easily tell the model's performance.

[jiashu-zhu]

Hi, have you get the answer?