Not match the pretrained ckpt about cmgan HOT 6 CLOSED

Can you show me your training loss tensorboard?

from cmgan.

Can you show me your training loss tensorboard?

Sorry I didn't record the loss on tensorboard, but we have use the early stop method since it gets overfitting when epoch number is too large, usually 80 epochs would be enough.

from cmgan.

I find that the --loss_weights", default=[0.1, 0.9, 0.2, 0.05] is not same as the paper default, and let loss_gan be smaller than the other loss, finally the GAN is not work

from cmgan.

Do you use extract dataset to train the pretrained ckpt?

from cmgan.

Since the GAN loss is much large than other losses, we add the weights to scale them to the same level. We didn't use any other dataset to train the model.

from cmgan.

Thanks your answer

from cmgan.