pjreddie / darknet

Convolutional Neural Networks
http://pjreddie.com/darknet/
Other
25.75k stars 21.33k forks source link

CUDA Error: unknown error #748

Open sezan92 opened 6 years ago

sezan92 commented 6 years ago

screenshot from 2018-04-30 17 09 20 Hello. My system is Ubuntu 14.04 65 bit architecture GPU NVIDA geforce 920mx Cuda 8.0 , cudnn 5.1

But everytime i run ./darknet imtest data/eagle.jpg after Gpu installation, this error comes

darknet: ./src/cuda.c:36: check_error: Assertion `0' failed.
Aborted (core dumped)

The result of nvidia-smi is attached

xhzzhang commented 6 years ago

I have same issue, dont know how to fix it.

xhzzhang commented 6 years ago

My message is : CUDA Error: an illegal memory access was encountered darknet: ./src/cuda.c:36: check_error: Assertion `0' failed. CUDA Error: an illegal memory access was encountered

ahsan856jalal commented 6 years ago

try ./darknet detector test data/coco.data cfg/yolo.cfg yolo.weights data/eagle.jpg Note: data,cfg and weight files can have different names and directories based on you Regards Ahsan

On Wed, Jun 20, 2018 at 3:44 PM Zhang Xu-Hui notifications@github.com wrote:

I have same issue, dont know how to fix it.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pjreddie/darknet/issues/748#issuecomment-398705982, or mute the thread https://github.com/notifications/unsubscribe-auth/AK9zjgxyX0B61gQi4mrRm3C3d7FaGsRWks5t-idygaJpZM4Tsbhw .

ahsan856jalal commented 6 years ago

you can also specify GPU ./darknet detector test data/coco.data cfg/yolo.cfg yolo.weights data/eagle.jpg -i 0[1 or 2]

Regards Ahsan

On Wed, Jun 20, 2018 at 3:45 PM Ahsan Jalal ahsan856jalal@gmail.com wrote:

try ./darknet detector test data/coco.data cfg/yolo.cfg yolo.weights data/eagle.jpg Note: data,cfg and weight files can have different names and directories based on you Regards Ahsan

On Wed, Jun 20, 2018 at 3:44 PM Zhang Xu-Hui notifications@github.com wrote:

I have same issue, dont know how to fix it.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pjreddie/darknet/issues/748#issuecomment-398705982, or mute the thread https://github.com/notifications/unsubscribe-auth/AK9zjgxyX0B61gQi4mrRm3C3d7FaGsRWks5t-idygaJpZM4Tsbhw .

ciwei123 commented 6 years ago

@sezan92 ,Hi, have you fix that error? I have same issue, dont know how to fix it. Thanks~

sezan92 commented 6 years ago

@ciwei123 no i couldnt , i had to run on another machine

ciwei123 commented 6 years ago

@ciwei123 no i couldnt , i had to run on another machine

Is the computer configuration causing an error?Could you tell me that machine's configuration?Thank you very much!

darthdeus commented 5 years ago

Same problem for me

layer     filters    size              input                output
    0 CUDA Error: unknown error
darknet: ./src/cuda.c:36: check_error: Assertion `0' failed.
[1]    23456 abort (core dumped)  ./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg
akshansh0074 commented 5 years ago

12 darknet: ./src/cuda.c:36: check_error: Assertion `0' failed. yolov3 CUDA Error: out of memory I am facing the same problem

darthdeus commented 5 years ago

@AKSHANSH47 It doesn't look like the same error. Yours says out of memory and not unknown error. You're probably just running out of GPU RAM.

MedAliMimouni commented 3 years ago

Guy for me I restarted my computer and it worked fine. if the error still, you can rebuild the darknet on your computer.