This project showcases a simple Generative Adversarial Network (GAN) implemented with PyTorch and trained on the MNIST dataset. The objective is for the GAN to reproduce the handwritten digits found in the MNIST dataset as accurately as possible.
- Generative Adversarial Network (GAN): Utilizes a generator and discriminator to reproduce the MNIST digits.
- Extensive Training: The model underwent almost 10,000 epochs to achieve satisfactory results.
After nearly 10,000 epochs of training, the GAN managed to produce images that closely resemble the handwritten digits from the MNIST dataset. You can view the generated images in the output/
directory.
- Python 3.x
- PyTorch
- Implement further optimization techniques to reduce the number of epochs required.
- Explore different GAN architectures for improved accuracy and efficiency.
Feel free to provide feedback or contribute to the project. All suggestions and contributions are welcome!