Not using Epoch but Iteration to train

yunjey / stargan

StarGAN - Official PyTorch Implementation (CVPR 2018)

MIT License

5.21k stars 967 forks source link

Not using Epoch but Iteration to train #126

Closed tranctan closed 4 years ago

tranctan commented 4 years ago

Hi, after looking into the code carefully in solver.py, I do not see the author use epoch to train, but merely number of iterations. This is not a usual case, but may have some reasonable effect for the model performance.

Can anyone explain why we don't particularly use Epoch to train in this case ? Thank you a lot !

tranctan commented 4 years ago

I'm closing this as this is answered here.

https://github.com/clovaai/stargan-v2/issues/47