search

chuanqi305 / MobileNet-SSD

Caffe implementation of Google MobileNet SSD detection network, with pretrained weights on VOC0712 and mAP=0.727.
MIT License
2k stars 1.18k forks source link

train.sh #159

Closed oolloosky closed 5 years ago

oolloosky commented 5 years ago

When I train with train.sh, it doesn't seem to report an error and then it ends. The information is probably like this (there is too much in front, I just intercepted the last part). I don't know why?

I0327 10:54:49.379165 5846 net.cpp:434] conv11/dw/relu <- conv11/dw I0327 10:54:49.379170 5846 net.cpp:395] conv11/dw/relu -> conv11/dw (in-place) I0327 10:54:49.379175 5846 net.cpp:150] Setting up conv11/dw/relu I0327 10:54:49.379179 5846 net.cpp:157] Top shape: 24 512 19 19 (4435968) I0327 10:54:49.379182 5846 net.cpp:165] Memory required for data: 3384791168 I0327 10:54:49.379185 5846 layer_factory.hpp:77] Creating layer conv11 I0327 10:54:49.379192 5846 net.cpp:100] Creating Layer conv11 I0327 10:54:49.379195 5846 net.cpp:434] conv11 <- conv11/dw I0327 10:54:49.379200 5846 net.cpp:408] conv11 -> conv11 @ 0x7ff7a30655d5 (unknown) @ 0x7ff7a0aa16ba start_thread @ 0x7ff7b035841d clone @ 0x0 (unknown) Aborted (core dumped)

oolloosky commented 5 years ago

Who can help me? thx

oolloosky commented 5 years ago

I found the error, before this information, because it didn't read the image correctly.