Software compressive optimization of deep neural networks

Software compressive optimization techniques are based on the need of deploying complex and large deep neural network models in devices with low processing capacity. Their high efficiencies allow deep neural networks to reduce the cost of resources while maintaining a good performance and reliable results. In this project, two optimization techniques are studied, analysed and combined in order to maximize the compression of a state-of-the-art deep neural network model. The results achieved show the different approaches that can be followed and the great impact that these techniques have in the downsizing of large models.