search

alexgkendall / SegNet-Tutorial

Files for a tutorial to train SegNet for road scenes using the CamVid dataset
http://mi.eng.cam.ac.uk/projects/segnet/tutorial.html
847 stars 517 forks source link

compute_bn_statistics.py runs error #93

Closed Jinming-Su closed 7 years ago

Jinming-Su commented 7 years ago

There is a strange error when I run compute_bn_statistics using my data whose resolution is 480 X 240.
Sometimes , it is 

I0426 11:20:20.092542  7266 net.cpp:235] This network produces output conv5_3-var
I0426 11:20:20.092551  7266 net.cpp:235] This network produces output conv5_3_D-mean
I0426 11:20:20.092559  7266 net.cpp:235] This network produces output conv5_3_D-var
I0426 11:20:20.092566  7266 net.cpp:235] This network produces output loss
I0426 11:20:20.092576  7266 net.cpp:235] This network produces output per_class_accuracy
I0426 11:20:20.092663  7266 net.cpp:482] Collecting Learning Rate and Weight Decay.
I0426 11:20:20.092689  7266 net.cpp:247] Network initialization done.
I0426 11:20:20.092700  7266 net.cpp:248] Memory required for data: 2930874452
61.0591571331
start rest:
start 1
80.4547159672
progress: 1/746
start 2
42.5934720039
progress: 2/746
start 3
42.8884651661
progress: 3/746
start 4
43.0134861469
progress: 4/746
start 5
42.8479809761
progress: 5/746
start 6
[1]    7266 segmentation fault (core dumped)  python compute_bn_statistics.py Models/segnet_train.prototxt  Inference

And sometime is, 

I0426 11:18:24.229485  7225 net.cpp:235] This network produces output conv4_3_D-var
I0426 11:18:24.229496  7225 net.cpp:235] This network produces output conv5_1-mean
I0426 11:18:24.229507  7225 net.cpp:235] This network produces output conv5_1-var
I0426 11:18:24.229519  7225 net.cpp:235] This network produces output conv5_1_D-mean
I0426 11:18:24.229532  7225 net.cpp:235] This network produces output conv5_1_D-var
I0426 11:18:24.229543  7225 net.cpp:235] This network produces output conv5_2-mean
I0426 11:18:24.229555  7225 net.cpp:235] This network produces output conv5_2-var
I0426 11:18:24.229568  7225 net.cpp:235] This network produces output conv5_2_D-mean
I0426 11:18:24.229579  7225 net.cpp:235] This network produces output conv5_2_D-var
I0426 11:18:24.229591  7225 net.cpp:235] This network produces output conv5_3-mean
I0426 11:18:24.229604  7225 net.cpp:235] This network produces output conv5_3-var
I0426 11:18:24.229614  7225 net.cpp:235] This network produces output conv5_3_D-mean
I0426 11:18:24.229625  7225 net.cpp:235] This network produces output conv5_3_D-var
I0426 11:18:24.229636  7225 net.cpp:235] This network produces output loss
I0426 11:18:24.229648  7225 net.cpp:235] This network produces output per_class_accuracy
I0426 11:18:24.229768  7225 net.cpp:482] Collecting Learning Rate and Weight Decay.
I0426 11:18:24.229805  7225 net.cpp:247] Network initialization done.
I0426 11:18:24.229818  7225 net.cpp:248] Memory required for data: 2930874452
65.9182889462
start rest:
start 1
[1]    7225 bus error (core dumped)  python compute_bn_statistics.py Models/segnet_train.prototxt  Inference

Please help!

Jinming-Su commented 7 years ago

I have solved by this https://github.com/alexgkendall/caffe-segnet/issues/62.