The spread of the disease caused by COVID 19 has changed the way people function and has given new obligations that people must adhere to in order to protect themselves and others. This includes keeping a distance, as well as the mandatory wearing of a protective face mask indoors. This paper explains the development and training of a model for object detection of wearing face masks and a comparison with a different configuration of the same model and a comparison with another model. In the past, creating a custom object detector required large amounts of resources, people and time. Reliable models can now be created faster and more efficiently with the TensorFlow 2 Object Detection API. This paper will explain the following: installation of the TensorFlow Object Detection API on Google Colab, installation of all necessary tools and libraries needed for accurate model operation, image preparation (to be performed locally using the LabelImg tool), model training with tracking graphs on TensorBoard, testing 3 model configurations (using images) on Google Colab, and finally comparing and explaining the results. The metrics used in the paper are also explained, as well as the convolutional networks and other terminology needed to understand the topic.