The aim of the bachelor thesis is to verify the accuracy of the system called Punctuator 2, which performs automatic punctuation restoration and sentence boundary detection, in the Slovak language. Verification included performing experiments using various parameters. Based on the evaluation and comparison of the results, it is possible to estimate suitable parameters for further work with the text in the Slovak language. The thesis also includes a theoretical analysis of the issue, which contains of basic information about neural networks, a more detailed description of the LSTM and GRU networks and characterization of the parameters precision, recall and F-score, by which it is possible to evaluate and compare the results of the experiments. Summary the current state of punctuation restoration issue using literature and the introduction of the Punctuator 2 model is also one of the parts of the bachelor thesis.