YOLOv8-ONNX-RKNN-HORIZON-TensorRT-Segmentation icon indicating copy to clipboard operation
YOLOv8-ONNX-RKNN-HORIZON-TensorRT-Segmentation copied to clipboard

Training the yolov8 neural network

Open lera5648 opened this issue 1 year ago • 1 comments

Good afternoon I tried converting rknn to a .pt model trained on the original ultralitics repository. But after converting the model to rknn, there were a large number of false and incorrect predictions (this was observed on the rockchip board. On the intel processor everything worked fine). Tell me, is there any special technique for training yolov8, in which the model is successfully transformed?

lera5648 avatar Dec 21 '23 04:12 lera5648

In the Intel processor, you can only get the simulation result, which can be totally different from the rockchip result When you tried to implement your rknn model into rockchip, you need to consider the which model of rockchip and the post-process program carefully

laitathei avatar Dec 23 '23 07:12 laitathei