Technically a GAN trained on 850 ganyu image in kaggle + 7000 image of anime face dataset. Trained on 120epoch in total(60+60=120).
Nothing special really. What did you expect from 944k parameter image generation model. It's a Normal GAN with local backpropagation on 64x64 image(via resizing).
Feel free to use anyway. either to fine tuning it or whatever. If you want to ask according to this model just ask in discussion.
Still newbie here, Happy Learning.
Stay tuned, maybe.
Dunno...
Image Result based on epoch 60 in second training
recon_epoch_60
Image Result based on epoch 60 in Third training (Training on 14850 images | Pure Layer-Wise Detached Forward Pass)
