Unable to initialize network in test_segmentation.cpp

[rawnam]

Hi,

I am trying to run the segnet example using test_segmentation.cpp. The test_segmentation_camvid.py runs perfectly after training but in trying to run the same test in cpp gives the following error. Please suggest what could be going wrong here.

F0220 15:27:45.372921 11359 test_segmentation.cpp:86] Check failed: net_->num_in puts() == 1 (0 vs. 1) Network should have exactly one input.
*** Check failure stack trace: ***
@ 0x7fd253e285cd google::LogMessage::Fail()
@ 0x7fd253e2a433 google::LogMessage::SendToLog()
@ 0x7fd253e2815b google::LogMessage::Flush()
@ 0x7fd253e2ae1e google::LogMessageFatal::~LogMessageFatal()
@ 0x40688c Classifier::Classifier()
@ 0x403813 main
@ 0x7fd2532d7830 __libc_start_main
@ 0x403d79 start
@ (nil) (unknown)
./test_segnet.sh: line 2: 11359 Aborted (core dumped) ./SegNet/caffe-segnet-cudnn5/build/examples/SegNet_with_C++/test_segmentation.bin ./SegNet/Models/segnet_inference.prototxt ./SegNet/Models/Inference/test_weights_segnet camvid.caffemodel ./SegNet/CamVid/0001TP_008550.jpg ./SegNet/Scripts/camvid11.png

Please note: I tried to change the input denseimagedata layer with simple caffe input layer and it worked but the output doesn't match with the camvid.py output. Probably a weight file mismatch, I think.