KWL
KWL
Hi, did you convert this model to onnx successfully?
Try to remove "/" before Example_Models/segnet_weights_driving_webdemo.caffemodel Same thing to Scripts/camvid12.png
I am also new to caffe but as far as I know, it is not strange that CPU takes longer time. Maybe you can reduce the batch_size in training prototxt.
#5 Look at this issue.
It is a picture I believe. In the folder SegNet-Tutorial/Scripts/ , there are two images that have colors - camvid12.png and camvid11.png, which are corresponding 12 classes and 11 classes.