Hello NXP,
I am running YOLOv8n with TFLite weights on i.xm8qm (version L5.10.72-2.2.0). The inference using the GPU delegate for yolov8n_float32.tflite and yolov8n_integer_quant.tflite works correctly. However, when I switch to using the VX delegate or the NNAPI delegate, the outputs are not the same,it is strange.
Please help me fix this issue.