The goal is to compare existing optimizers for neural networks and understand which are most suitable for practical ML4SE tasks. We are evaluating several optimizers for a number of implemented models for the method name prediction task.