Gun violence is a severe problem in the world, particularly in the UnitedStates. Computer vision methods have been studied to detect guns insurveillance video cameras or smart IP cameras and to send a real-time alert tosafety personals. However, due to no public datasets, it is hard to benchmarkhow well such methods work in real applications. In this paper we publish adataset with 51K annotated gun images for gun detection and other 51K croppedgun chip images for gun classification we collect from a few different sources.To our knowledge, this is the largest dataset for the study of gun detection.This dataset can be downloaded at www.linksprite.com/gun-detection-datasets. Wealso study to search for solutions for gun detection in embedded edge device(camera) and a gun/non-gun classification on a cloud server. This edge/cloudframework makes possible the deployment of gun detection in the real world.