Producer process has been terminated before all shared CUDA tensors released

[garylidd]

#7 #6

[hey-it-s-me]

Hey~Very appreciate to your great work! I also meet the problem of the 'producer process terminated', I wanna know when will the problem be fixed?

[garylidd]

Hi, we have tried some methods and still cannot eliminate this warning, but this should only appear after the program ends and does not affect the result.

[Yuhanelle]

Same error here but I got the warning before the program ends, which returns no results.
I got "No such file or directory" error when the kf_buffer.get() is called at the end of the program. Has this issue been shown before? Thank you in advance.

[moyutianque]

Same error here but I got the warning before the program ends, which returns no results. I got "No such file or directory" error when the kf_buffer.get() is called at the end of the program. Has this issue been shown before? Thank you in advance.

I have encountered the same issue, have you solved it? It seems "share_data.stop_mapping" is not correctly locked, the value sometimes changed from False to True in the last iteration of mapping.py int the while loop, just around kf_buffer.get(). In addition, when stop_mapping is set, the kf_buffer is not empty, which will make another round of do_mapping. Currently I simple skip the get() which causes error by try-except.

[random-guest]

I got the same error, with no results at the end.

@garylidd any updates on this?

[Yuhanelle]

Same error here but I got the warning before the program ends, which returns no results. I got "No such file or directory" error when the kf_buffer.get() is called at the end of the program. Has this issue been shown before? Thank you in advance.

I have encountered the same issue, have you solved it? It seems "share_data.stop_mapping" is not correctly locked, the value sometimes changed from False to True in the last iteration of mapping.py int the while loop, just around kf_buffer.get(). In addition, when stop_mapping is set, the kf_buffer is not empty, which will make another round of do_mapping. Currently I simple skip the get() which causes error by try-except.

if you move this to the top of the while loop in spin() mapping.py, it solves the issue.
if share_data.stop_mapping:
break
elif not kf_buffer.empty():
tracked_frame = kf_buffer.get()
# self.create_voxels(tracked_frame)
...

[random-guest]

@Yuhanelle thank you so much.

[HelloTaeyeon]

@Yuhanelle thank you so much.

Have you solved this problem yet？