yolov8在板端运行时，报错slice operator axes' size must be the same ........

我复现论坛yolov8那篇文章，模型转换过程很顺利没有报错

但是使用文章里面的部署文件时发生报错

报错信息如下

[HBRT] set log level as 0. version = 3.14.5

[DNN] Runtime version = 1.9.7_(3.14.5 HBRT)

tensor type: NV12_SEPARATE

data type: uint8

layout: NCHW

shape: (1, 3, 416, 416)

[E][DNN][slice.cpp:90](1679747511866) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747511867) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747511996) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747511997) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512120) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512120) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512233) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512233) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512347) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512347) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512464) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512464) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512578) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512578) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512694) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512694) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512812) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512813) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512929) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747512929) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747514520) slice operator axes' size must be the same with input tensor dim size when missing steps

[E][DNN][slice.cpp:90](1679747514520) slice operator axes' size must be the same with input tensor dim size when missing steps

Traceback (most recent call last):

File "inference.py", line 214, in <module>

results = nms(*results)

File "inference.py", line 157, in nms

indices = cv2.dnn.NMSBoxes(boxes, confidences, score_thres, iou_thres).flatten()

AttributeError: 'tuple' object has no attribute 'flatten'

【INFO】: Offload model "ppyolo_tiny_416x416_nv12" Successfully.