Data labeling is an essential setup preparing data for computer vision applications, especially for object detection and segmentation due to several reasons: time-consuming, mislabeling, large-scale data, imbalances between classes, and expertise with object categories. Addressing these challenges and finding efficient labeling strategies is crucial for advancing object detection and computer vision research. The aim of this work is to use a self-supervised learning model, the Grounding Distillation-Based Image Navigation and Orientation (DINO) that belongs to Meta company to auto-annotate our data. The DINO model uses the knowledge of self-supervised learning to automatically generate labels for our objects in the dataset for object detection tasks. This self-supervised model reduced the time and complexity of labeling big-scale datasets and that will help to create a dataset for object detection and other computer vision tasks faster and easier. In the results section of this paper, we present how the DINO model could generate auto-labeling data for [Cars, Buses, Motorcycles, and Persons].