Since the first discovery of the binary black hole merger, the gravitational wave detectors, Advanced LIGO, Advanced Virgo, and KAGRA have provided astonishing information about our universe. Persistent, quasi-monochromatic gravitational waves, so-called continuous gravitational waves are also targets of these detectors. We expect continuous waves to bring information about various aspects of physics such as the inner structure of the neutron stars, the existence of a boson cloud around a rotating black hole, and the abundance of the sub-solar mass primordial black holes. Searches for continuous gravitational waves are the most computationally expensive task in gravitational wave astronomy. Therefore, deep learning application is actively studied. In this talk, I will briefly review the continuous gravitational waves and introduce three approaches to applying deep learning for continuous wave searches.