Segmentation Fault with multiprocessing (num_workers>0)

google-ai-edge / mediapipe

Cross-platform, customizable ML solutions for live and streaming media.

https://ai.google.dev/edge/mediapipe

Apache License 2.0

27.81k stars 5.18k forks source link

Segmentation Fault with multiprocessing (num_workers>0) #3194

Closed mdv3101 closed 1 year ago

mdv3101 commented 2 years ago

System information

Have I written custom code (as opposed to using a stock example script provided in MediaPipe): Yes
OS Platform and Distribution (e.g., Linux Ubuntu 16.04, Android 11, iOS 14.4): CentOS Linux release 7.9.2009, cuda10.2, cudnn7.6.5
Programming Language and version ( e.g. C++, Python, Java): Python 3.7.1
MediaPipe version: 0.8.7.1
Solution ( e.g. FaceMesh, Pose, Holistic ): FaceMesh

Describe the current behavior: I am using mediapipe to get face landmarks in a dataloader using PyTorch. The code works fine without multiprocessing. But when I increase the number of workers to 1(or more), the code crashes and shows segmentation fault. I have installed mediapipe using pip.

sureshdagooglecom commented 2 years ago

Hi @mdv3101 , did you get a chance to check this closed issue https://github.com/google/mediapipe/issues/1720

mdv3101 commented 2 years ago

Hi @sureshdagooglecom Yes, I checked that issue. Apparently, the error is not solved in that as well.

google-ml-butler[bot] commented 2 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you.

google-ml-butler[bot] commented 2 years ago

Closing as stale. Please reopen if you'd like to work on this further.

google-ml-butler[bot] commented 2 years ago

Are you satisfied with the resolution of your issue? Yes No

mdv3101 commented 2 years ago

Hi @sureshdagooglecom Can provide any solution to this issue?

mdv3101 commented 2 years ago

Hi @ivan-grishchenko , Can you look into this issue?

kuaashish commented 1 year ago

Hi @mdv3101, Would you please let us know this is still an issue or has been resolved in the latest version 0.8.11? If issue persists could you share the complete steps to reproduce the same from our end.

google-ml-butler[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you.

google-ml-butler[bot] commented 1 year ago

Closing as stale. Please reopen if you'd like to work on this further.

google-ml-butler[bot] commented 1 year ago

Are you satisfied with the resolution of your issue? Yes No

milely commented 1 year ago

Hello, have you solved this problem? I use mediapipe in pytorch's Dataset, when the num_workers of Dataloader > 0, the program will get stuck.

Daniel-Kelvich commented 1 year ago

I encountered the same issue.

crj1998 commented 4 months ago

same issue here