I trained a DCGAN model on a dataset. Few things have been observed:
- Repetitive patterns in discriminator loss (d_loss = d_loss_real + d_loss_fake) and g_loss
- d_loss_real (discriminator loss on real img) and d_loss_fake (discriminator loss on fake img) are negatively correlated
I ran the model for 6000 epochs (24 batches of data)
Question: At which point should I stop the training? Do I need to run it for few more epochs? or is it a problem with the data? (I can go with WGANs but what can I improve if I’m using DCGAN?