This ponder centers on the discovery of weapons in pictures employing a profound learning demonstrate based on the YOLO (You Merely See Once) system. The essential objective is to prepare a custom weapon location show employing a dataset from the Roboflow stage and fine-tune it to distinguish different sorts of weapons in genuine time. The venture utilizes YOLOv5, a well established question location demonstrate known for its speed and exactness in distinguishing objects inside pictures. The workflow starts with downloading and planning a dataset, particularly the "Weapon Discovery" dataset, from Roboflow and setting it up for preparing. Utilizing the YOLOv5 system, the show is prepared on this dataset with a arrangement custom-made to the issue of weapon location. Once prepared, the demonstrate is assessed for execution utilizing approval information, and forecasts are made on modern pictures containing potential weapon objects. Bounding boxes are drawn around recognized weapons, with a certainty score showing the model's certainty almost each forecast. The comes about are visualized utilizing Python's Matplotlib library to show the pictures nearby their predicted bounding boxes and course names. The demonstrate gives a powerful instrument for computerized weapon location, valuable for security frameworks, reconnaissance, and other related applications. By leveraging both Roboflow and YOLOv5, this extend illustrates a viable approach to tackling genuine world issues including question discovery, exhibiting the potential of profound learning strategies for moving forward security and security.