"Bus error: 10" when training

[Matthias Macé]

Hi

I am trying to train CRFasRNN with my own data on a mac using https://github.com/martinkersner/train-CRF-RNN

I am able to produce lmdb using 500px (largest edge) but when i make "python2.7 ", I am getting a "Bus error: 10".

What am I doing wrong please ?

Here is the end of the log :

I1126 22:08:00.789855 3454440384 net.cpp:164] Top shape: 1 2 500 500 (500000)

I1126 22:08:00.789860 3454440384 layer_factory.hpp:76] Creating layer splitting

I1126 22:08:00.789866 3454440384 net.cpp:111] Creating Layer splitting

I1126 22:08:00.789870 3454440384 net.cpp:478] splitting <- coarse

I1126 22:08:00.789875 3454440384 net.cpp:434] splitting -> unary

I1126 22:08:00.789880 3454440384 net.cpp:434] splitting -> Q0

I1126 22:08:00.789885 3454440384 net.cpp:156] Setting up splitting

I1126 22:08:00.789888 3454440384 net.cpp:164] Top shape: 1 2 500 500 (500000)

I1126 22:08:00.789892 3454440384 net.cpp:164] Top shape: 1 2 500 500 (500000)

I1126 22:08:00.789897 3454440384 layer_factory.hpp:76] Creating layer inference1-ft

I1126 22:08:00.789903 3454440384 net.cpp:111] Creating Layer inference1-ft

I1126 22:08:00.789907 3454440384 net.cpp:478] inference1-ft <- unary

I1126 22:08:00.789911 3454440384 net.cpp:478] inference1-ft <- Q0

I1126 22:08:00.789916 3454440384 net.cpp:478] inference1-ft <- data_data_0_split_2

I1126 22:08:00.789921 3454440384 net.cpp:434] inference1-ft -> pred

I1126 22:08:00.789930 3454440384 multi_stage_meanfield.cpp:49] This implementation has not been tested batch size > 1.

Bus error: 10