Segmentation fault (core dumped) with some pretrained models (but not all) in creating my own skeleton data

[suijua]

Hello,

I have a strange issue when loading some pretrained models. On loading, some models abort and give an "Segmentation fault (core dumped)" message, like below:

[kennymckormick]

Do not directly use this python script, follow the usage in the ipython notebook and use dist_run.sh to launch the job.

[suijua]

Do not directly use this python script, follow the usage in the ipython notebook and use dist_run.sh to launch the job.

Thanks for your reply ! I ran the above mentioned link, but it's something wrong just as this:

[kennymckormick]

Hi, I'm sorry, I cannot figure out what might be the problem. I haven't met that problem on every linux GPU server I used.

[suijua]

Hi, I'm sorry, I cannot figure out what might be the problem. I haven't met that problem on every linux GPU server I used.

Thank you for your reply! that is all right

[yuchen-ji]

Do not directly use this python script, follow the usage in the ipython notebook and use dist_run.sh to launch the job.

Thanks for your reply ! I ran the above mentioned link, but it's something wrong just as this:

I have the same problem, have you fixed it?

[azelee]

I have tried to change the parameter device of det_model and pose_model, and of course import torch is required! It works in term of my code and env

[AgileChen3]

I have tried to change the parameter device of det_model and pose_model, and of course import torch is required! It works in term of my code and env

I have the same problem, Can you share your code with me， thank you

[azelee]

I have tried to change the parameter device of det_model and pose_model, and of course import torch is required! It works in term of my code and env

I have the same problem, Can you share your code with me， thank you

I just changed the det_model and pose_model as my notes above and insert import torch as well, It worked in my server!

[samouha]

Hi, @yuchen-ji have you, please, solved this issue? I have the same segmentation fault problem! And the proposed solution by @azelee doesn't work unfortunately :/

[kennymckormick]

Recently I also met this problem. I guess the potential reason be a new version of gcc is used to compile the open-mmlab codebases, which lead to some errors. Now I have fixed it by using a very specific conda environment for this project. Please following the new installation guide to reinstall pyskl and see if the problem has been fixed now. Sorry for the late fix.

[kennymckormick]

Feel free to re-open the issue if the problem still exists when you install following the latest installation guide.

[suijua]

Feel free to re-open the issue if the problem still exists when you install following the latest installation guide.

ok, thanks for your work!

[JiaweiMorris]

Recently I also met this problem. I guess the potential reason be a new version of gcc is used to compile the open-mmlab codebases, which lead to some errors. Now I have fixed it by using a very specific conda environment for this project. Please following the new installation guide to reinstall pyskl and see if the problem has been fixed now. Sorry for the late fix.

hi, I met the same problem again, and I followed the new installation guide or @azelee's solution but they didn't work.

[seekFire]

@JiaweiMorris I also met recently