lewes6369 / TensorRT-Yolov3

TensorRT for Yolov3
MIT License
489 stars 165 forks source link

ERROR: Cuda initialization failure with error 35. Please check cuda installation #61

Closed ZhangYouYu-09 closed 4 years ago

ZhangYouYu-09 commented 5 years ago

Hello @lewes6369 .

cuda9.2, NVIDIA-SMI 384.130,
i run the Yolov3 and some issue occured

user@user-S2600WT2R:~/youyu/TensorRT-Yolov3-Github$ ./install/runYolov3 --caffemodel=v3face/yolov3-widerface_final.caffemodel --prototxt=v3face/yolov3_608_trt.prototxt --input=test.jpg --W=608 --H=608 --class=1 ####### input args####### C=3; H=608; W=608; caffemodel=v3face/yolov3-widerface_final.caffemodel; calib=; class=1; evallist=; input=test.jpg; mode=fp32; nms=0.450000; outputs=yolo-det; prototxt=v3face/yolov3_608_trt.prototxt; ####### end args####### init plugin proto: v3face/yolov3_608_trt.prototxt caffemodel: v3face/yolov3-widerface_final.caffemodel ERROR: Cuda initialization failure with error 35. Please check cuda installation: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html. Segmentation fault (core dumped)

lewes6369 commented 5 years ago

as the info shows, "Cuda initialization failure with error 35. Please check cuda installation".And can you run other the official tensorRT samples? It should be failed, too.

ZhangYouYu-09 commented 5 years ago

i got it. :) thanks so much


From: lewes6369 notifications@github.com Sent: Sunday, September 8, 2019 2:53:22 PM To: lewes6369/TensorRT-Yolov3 TensorRT-Yolov3@noreply.github.com Cc: ZhangYouYu-09 ZyyGlenn@Outlook.com; Manual manual@noreply.github.com Subject: Re: [lewes6369/TensorRT-Yolov3] ERROR: Cuda initialization failure with error 35. Please check cuda installation (#61)

as the info shows, "Cuda initialization failure with error 35. Please check cuda installation".And can you run other the official tensorRT samples? It should be failed, too".

― You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/lewes6369/TensorRT-Yolov3/issues/61?email_source=notifications&email_token=AMH57Z4ZE2SCLIFJYR7LRS3QISOOFA5CNFSM4IL6Q5W2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6FJIKI#issuecomment-529175593, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AMH57Z7NJCBJ3Q7BDIIBRQLQISOOFANCNFSM4IL6Q5WQ.

1LOVESJohnny commented 4 years ago

I met the same problem. In my case, it's caused by mismatched CUDA versions between my local environment CUDA 10.1 and within a specific docker CUDA 10.2. I updated my local environment to CUDA 10.2 and then the problem is gone.

tangsipeng commented 4 years ago

update driver of CUDA 10.2 then done

zhangxiaoya commented 3 years ago

may be your Nvidia Driver failed, try run nvidia-smi, if get error message, re-install nvidia driver can fix it.