YOLOv4 is a state-of-the-art object detection model that achieves high accuracy and real-time performance. It combines various advanced features such as Weighted-Residual-Connections (WRC), Cross-Stage-Partial-connections (CSP), Cross-mini-Batch Normalization (CmBN), Self-adversarial-training (SAT), and Mish-activation. It also incorporates Mosaic data augmentation, DropBlock regularization, and CIoU loss. YOLOv4 achieves a 43.5% AP (65.7% AP50) on the MS COCO dataset at a real-time speed of ~65 FPS on a Tesla V100 GPU. The model is designed to be efficient and suitable for single GPU training, making it accessible for widespread use. It improves upon previous models like YOLOv3 by enhancing accuracy and speed through a combination of novel techniques and modifications to existing methods. The model's architecture includes a CSPDarknet53 backbone, SPP and PANet for feature aggregation, and a YOLOv3 head. YOLOv4 also incorporates a variety of "Bag of Freebies" and "Bag of Specials" techniques to enhance performance. These include data augmentation methods like Mosaic and CutMix, regularization techniques like DropBlock, and advanced loss functions like CIoU. The model is trained using a single GPU, making it highly efficient and accessible. YOLOv4 demonstrates superior performance compared to other state-of-the-art object detectors in terms of both speed and accuracy.YOLOv4 is a state-of-the-art object detection model that achieves high accuracy and real-time performance. It combines various advanced features such as Weighted-Residual-Connections (WRC), Cross-Stage-Partial-connections (CSP), Cross-mini-Batch Normalization (CmBN), Self-adversarial-training (SAT), and Mish-activation. It also incorporates Mosaic data augmentation, DropBlock regularization, and CIoU loss. YOLOv4 achieves a 43.5% AP (65.7% AP50) on the MS COCO dataset at a real-time speed of ~65 FPS on a Tesla V100 GPU. The model is designed to be efficient and suitable for single GPU training, making it accessible for widespread use. It improves upon previous models like YOLOv3 by enhancing accuracy and speed through a combination of novel techniques and modifications to existing methods. The model's architecture includes a CSPDarknet53 backbone, SPP and PANet for feature aggregation, and a YOLOv3 head. YOLOv4 also incorporates a variety of "Bag of Freebies" and "Bag of Specials" techniques to enhance performance. These include data augmentation methods like Mosaic and CutMix, regularization techniques like DropBlock, and advanced loss functions like CIoU. The model is trained using a single GPU, making it highly efficient and accessible. YOLOv4 demonstrates superior performance compared to other state-of-the-art object detectors in terms of both speed and accuracy.