Generative Adversarial Network

This repo is about using the Generative Adversarial Network (GAN) model for creating new faces with using CelebA dataset.

References

• Self-Attention Generative Adversarial Networks
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Information-theoretic stochastic contrastive conditional GAN: InfoSCC-GAN
• Instance Normalization: The Missing Ingredient for Fast Stylization
• Spectral Normalization for Generative Adversarial Networks
• Large-scale CelebFaces Attributes (CelebA) Dataset

Initial

Results

---

Frameworks: Tensorflow, WanDB, NumPy, Matplotlib