This repository contains the assignments for the course Deep Learning in Data Science at KTH Royal Institute of Technology
Assignment 1: In this assignment we built an one-layer network to classify images from the CIFAR-10 dataset. We trained the model using the Mini-Batch Gradient Descent.
Assignment 2: In this assignment it has been built a two-layer network to classify images from the CIFAR-10 dataset. The model has been trained with the Mini-Batch Gradient Descent with cyclical learning rates.
Assignment 3: In this assignment it has been built a multi-layer network to classify images from the CIFAR-10 dataset. The model has been trained with the Mini-Batch Gradient Descent with cyclical learning rates.
Assignment 4: We train a vanilla RNN to synthesize English text character by character, using the text from the book The Goblet of Fire by J.K. Rowling. The variation of SGD that is utilised for the optimization is the 'AdaGrad'.