This thesis examines the problematics of classification using artificial neural networks within use of the Keras framework, a high-level deep learning API for the Python programming language. The aim of the thesis is to discover the diverse options Keras has to offer in the field of classification, and to compare different types and topologies of artificial neural networks in the form of experiments on selected datasets, complemented by a simple experimental application whose main purpose is to provide an interface for these experiments.
- xpysik00-thesis.pdf - The main technical report.
- /text - Source code of the technical report.
- /app - The experimental application (including instructions and requirements).
- /experiments-results - Results of the experiments (plots and logs).
- /model-plots - Plots of model architectures used in the conducted experiments.