Using deep learning DCGAN neural network to generate realistic faces from celebrity images.
Using a Deep Convolutional Generative Adversarial Network (DCGAN) to generates new images of faces. The dataset that's used is the CelebFaces Atrribute Dataset (CelebA), which contains over 200,000 celebrity faces with annotations.
Tensorflow implementation of Deep Convolutional Generative Adversarial Networks which is a stabilize Generative Adversarial Networks.
| Hyperparameter | Number |
|---|---|
| Epochs | 1 |
| Batch size | 128 |
| Learning rate | 0.0005 |
| Z dimension | 100 |
| Beta 1 | 0.5 |
After 10 iterations:
After 20 iterations:
After 10 iterations:
After 20 iterations:
