如题

AI工程师你好，tusimple-onnx的尺寸为(320,800)，为什么项目例onnxruntime的推理detect部分先将图片resize到（400，800）然后裁剪从最后一行往上数320行，得到(320,800)的图片呢，直接resize到(320,800)会报错我试了，为什么这样呢，我先resize到其他比例可以吗，有什么原则吗

outnames :
exist_col
exist_row
loc_col
loc_row
[ERROR:[email protected]] global net_impl.cpp:1164 cv::dnn::dnn4_v20221220::Net::Impl::getLayerShapesRecursively OPENCV/DNN: [Reshape]:(onnx_node!Reshape_48): getMemoryShapes() throws exception. inputs=1 outputs=1/1 blobs=0
[ERROR:[email protected]] global net_impl.cpp:1167 cv::dnn::dnn4_v20221220::Net::Impl::getLayerShapesRecursively input[0] = [ 1 8 10 25 ]
[ERROR:[email protected]] global net_impl.cpp:1171 cv::dnn::dnn4_v20221220::Net::Impl::getLayerShapesRecursively output[0] = [ 1 4000 ]
[ERROR:[email protected]] global net_impl.cpp:1177 cv::dnn::dnn4_v20221220::Net::Impl::getLayerShapesRecursively Exception message: OpenCV(4.7.0) D:\tools\opencv\opencv-4.7.0\modules\dnn\src\layers\reshape_layer.cpp:150: error: (-1:Backtrace) Can't infer a dim denoted by -1 in function 'cv::dnn::computeShapeByReshapeMask'

OpenCV: terminate handler is called! The last OpenCV error is:
OpenCV(4.7.0) Error: Backtrace (Can't infer a dim denoted by -1) in cv::dnn::computeShapeByReshapeMask, file D:\tools\opencv\opencv-4.7.0\modules\dnn\src\layers\reshape_layer.cpp, line 150

大佬你好，感谢您的开源工作！我在使用你提供ufldv2_culane_res18_320x800.onnx文件在main.py推理时成功，但在使用我自己转换的culane_res18.onnx文件无法成功，出现如下报错：（culane_res18.onnx是利用官方权重转换而来）
/home/pc/anaconda3/envs/UFLD/bin/python3.7 /home/pc/FY/Ultra-Fast-Lane-Detection-v2-master/ONNX/opencv/main.py
Traceback (most recent call last):
File "/home/pc/FY/Ultra-Fast-Lane-Detection-v2-master/ONNX/opencv/main.py", line 99, in
net = Ultra_Fast_Lane_Detection_v2(args.modelpath)
File "/home/pc/FY/Ultra-Fast-Lane-Detection-v2-master/ONNX/opencv/main.py", line 10, in init
self.input_height = int(input_shape[0])
ValueError: invalid literal for int() with base 10: 'res18'