Sleep stage classification is one of the most important methods to diagnose narcolepsy and sleep disorders. By analyzing the polysomnogram, which includes bioelectrical signals such as EEG and ECG, the whole night’s sleep is divided into 30-second epochs, each belonging to five sleep stages: Wake, N1, N2, N3, and REM stages, according to the AASM guidelines. As deep learning has made breakthroughs in various fields in recent years, automatic sleep stages classification tasks are also undergoing a revolution from traditional methods to deep learning methods. Models combining convolutional neural networks and recurrent neural networks (e.g., LSTM) achieve state-of-the-art performance on many benchmark datasets.