Closed jianwu585218 closed 6 years ago
Try the latest dlib version.
Thanks for your reply.
I install the latest dlib version in the first . When i import the latest and torch in linux ,will get the 《Segmentation fault (core dumped) 》error.
I Install mang version , the dlib 19.8.1 with pytorch0.4 in other's machine is fine. So i want install dlib 19.8.1, but import dlib19.8.1,I get the
the face_recognition use the dlib
Ok, well, you need to fill out the issue template and follow its instructions if you expect to get any help. In particular, include exactly what I can type starting from a clean ubuntu install to reproduce this problem.
l get the same problem when I use cuda 8.0 and cudnn7 ,while use cuda8.0 and cudnn6 ,the error is: Error while calling cudnnActivationForward(context(), relu_activation_descriptor(), &alpha, descriptor(src), src.device(), &beta, descriptor(dest), dest.device()) in file /opt/dlib-master/dlib/cuda/cudnn_dlibapi.cpp:1465. code: 9, reason: CUDNN_STATUS_NOT_SUPPOR
I slove it but i don't why. you should import face_recognition before torch,such as:
demo.py:
import dlib
import torch
# other code
Segmentation fault (core dumped)
Please read the issue submission instructions and follow them to submit a bug report.
I happened the same problem when I used the face detection function of dlib. Also follow the instruction above. I move the import dlib front of import torch. Amazing, it worked. Hope someone can tell the reason.
@sunguwei I don't konw the reason.Please tell me if you find the reason ,thank u
I slove it but i don't why. you should import face_recognition before torch,such as:
same error , it works to me. Have you found the reason
Hi,I got a trouble that when i import torch and the lasted dlib in linux , it will get《Segmentation fault (core dumped) 》error. I found that install dlib-19.8.1 in win10 will be fine. So I try to install 19.8.1 in linux ,after I download the dlib-19.8.1.tra.gz and use command:pip install dlib-19.8.1.tra.gz, it shows successfully install But when i import dlib ,there is a new bug Please help me,Thank you very much