CS2 object detection with aimbot (YOLOv4-tiny)

So I've been looking into image processing and object detection particularly with the YOLO models and Darknet framework.

In this video it was using CPU for inference, which is why the detections is slow and not consistent. I've been trying to build OpenCV's DNN module with CUDA for quite some time now, but it doesn't seem to be that easy; via vcpkg - manually is nearly impossible due to errors upon errors.

Once I figure out how to utilize the GPU instead of CPU the inference will get much faster, approx. from 2-400 ms to 5-10ms - which means it will be more reliable for an aimbot/aim assistance.

Model was trained with custom dataset via Roboflow's annotation tools and export of the data and Darknet for the actual training. Currently with 339 images. maP%=0.7746, best=0.7763

Видео CS2 object detection with aimbot (YOLOv4-tiny) канала REVRBE