Skills: Python, Jupyter, PyTorch, GANs, Computer Vision, Scikit-Image, OpenCV, Pandas, Matplotlib.
Description: In this simple project, I experimented with the performance difference of Vanilla GAN, Least Squares GAN, and Deep Convolutional GAN when generating handwritten digits that resemble those from the MNIST dataset. The picture on the left is a collection of the images generated by the DCGAN, starting with random noise.
This was a part of the Stanford CS231n CNNs for Computer Vision course.