Deep learning tutorial error on kernel size, kernel_h and kernel_w

I’m new to Caffe, and I’m following Deep Learning Tutorial. I have created a python script to read the images, and create the databases (Steps 1 and 2). It works. Then I created another script to train and validate (steps 3-15) and also solver file. The script creates the networks (fcn_train and fcn_test, contents match tutorial), however the script fails at solver = caffe.SGDSolver(‘fcn_solver.prototxt”) The error message is below: I0213 01:30:33.690608 24127 layer_factory.hpp:76] Creating layer data I0213 01:30:33.691711 24127 net.cpp:111] Creating Layer data I0213 01:30:33.691727 24127 net.cpp:434] data -> data I0213 01:30:33.69281124379 db_lmdb.cpp:22] Opened lmdb train_images_lmdb/ ….. I0213 01:30:33.744316 24127 net.cpp:478] upscore <- score_classes I0213 01:30:33.744321 24127 net.cpp:434] upscore -> upscore F0213 01:30:33.744328 24127 base_conv_layer.cpp:19] Check failed: !conv_param.has_kernel_size() != !(conv_param.has_kernel_h() && conv_param.has_kernel_w()) Filter size is kernel_size OR kernel_h and kernel_w;…


Link to Full Article: Deep learning tutorial error on kernel size, kernel_h and kernel_w