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

error in running demo.py #153

Open meltingCat opened 5 years ago

meltingCat commented 5 years ago

F0113 17:47:26.589298 7077 cudnn_conv_layer.cpp:53] Check failed: status == CUDNN_STATUS_SUCCESS (4 vs. 0) CUDNN_STATUS_INTERNAL_ERROR Check failure stack trace: Aborted (core dumped)

using GTX1060 6G

rubmarsan commented 5 years ago

Same here, I can see in nvidia-smi that the script takes all available GPU Memory, don't really know why

runauto commented 5 years ago

same with me，my GPU is：1080ti

lingfengyueguang commented 5 years ago

How do you solve this problem at last?

Wallace00V commented 5 years ago

the same problem, my GPU is: Tesla P100

Wallace00V commented 5 years ago

uncomment engine:CAFFE in demo.py，the probloem is solved

jaysimon commented 5 years ago

uncomment engine:CAFFE in demo.py，the probloem is solved

can you tell us which line it is to uncomment?

dasol92 commented 5 years ago

uncomment engine:CAFFE in demo.py，the probloem is solved

can you tell us which line it is to uncomment?

@jaysimon uncommnet engine:CAFFE in deploy.prototxt at each layers.

nankingV commented 4 years ago

may that be your cudnn level too high for ssd?