The main objectives for this project were to use TensorFlow and Keras to implement a convolutional neural network capable of image classification and training a neural network that allows the identification of a non-linear and chaotic dynamic system.
To achieve this the neural network was trained with a stochastic gradient descent and a categorical cross-entropy loss. I was tasked with setting the value of the remaining hyper-parameters, performance metrics, and generating relevant figures and results for tuning and evaluation of the model.