Correct naming of YOLOv8 output layers

[maxsitt]

Output layers of the YOLOv8 ONNX model are currently named output1_yolov6r2, output2_yolov6r2, output3_yolov6r2.
https://github.com/luxonis/tools/blob/master/yolo/export_yolov8.py#L60

Not sure if this is on purpose for inference to work on the OAK. Correct naming of the output layers (output1_yolov8 etc.) might make things clearer for manual conversion afterwards.

[tersekmatija]

Hey, good catch! This is indeed intentional. Since both use the same processing, we accelerated the support for YoloV8 by naming them like YoloV6 layers. This will be fixed once support on FW (in our DepthAI) for YoloV8 directly is added with the new release.