EdgeFirst released a Model Zoo of YOLO object-detection and segmentation models, spanning YOLOv5 through YOLO26, validated across multiple edge hardware accelerators with more than 837 published validation sessions. Every published benchmark figure links back to the validation session that produced it, including model artifacts, dataset versions, timing traces, and hardware specifications. The post shows how measurement choices such as decoder selection can shift reported accuracy by more than 2 mAP, and offers EdgeFirst Studio and Profiler so users can reproduce the same validation workflow on their own datasets and hardware.
