Yolo Through Time Versions[1-8]: Arabic In-Depth Explanation

This course provides a complete journey through the YOLO family from YOLOv1 to YOLOv8, with clear and easy explanations delivered in Arabic. It focuses on understanding the internal architecture of each version, including backbone, neck, head, loss functions, and the key design improvements that made YOLO models more accurate, faster, and stronger over time.

You will start with YOLOv1 fundamentals, then explore the evolution in YOLOv2 and YOLOv3 such as anchor boxes and multi-scale detection. The course offers a detailed yet simple explanation of YOLOv4 components including CSPNet, FPN, PANet, SPP, Bag of Freebies, Bag of Specials, IoU-based losses, and evaluation metrics like Precision, Recall, and mAP.

It then moves to practical YOLOv5 topics such as data annotation, fine-tuning on custom datasets, multi-GPU training, speeding up inference with ONNX, TensorRT, and TFLite, and understanding key modules like C3 and SPPF. After that, you will study YOLOv6 architecture, anchor-free detection, label assignment strategies like SimOTA and Task-Aligned Learning, and important loss functions.

Finally, the course covers YOLOv7 architectural scaling and re-parameterization concepts, and concludes with a clear explanation of YOLOv8 architecture and repository structure. The entire course is explained in Arabic with an easy, step-by-step approach suitable for AI engineers and computer vision practitioners who want a deep yet simplified understanding of YOLO internals.