virtuerdem/gan-human-faces

GAN: Human Faces Generator

A Generative Adversarial Network trained to generate human faces.

Model Details

• Generator: Keras model (.keras)
• Discriminator: Keras model (.keras)
• Framework: TensorFlow/Keras
• Input dimension: 100 (latent space)
• Output: 64x64 RGB images

Usage

import tensorflow as tf

# Model yükleyin
gen = tf.keras.models.load_model('generator.keras')

# Rastgele latent vector oluşturun
import numpy as np
z = np.random.normal(0, 1, (1, 100))

# Face generate edin
fake_face = gen.predict(z)

Training Details

• Trained on: kaggle~~
• Epochs: 1000
• Batch size: 256