-
-
Notifications
You must be signed in to change notification settings - Fork 6
RuntimeError: unable to open shared memory object </torch_2276740_2849291446> in read-write mode #158
Comments
The same thing happened again although, interestingly, this time the script created a bunch more batches before finally dying... This feels like an upstream bug in PyTorch. Maybe the ultimate solution for us is #86
|
I just did a quick search of the PyTorch GitHub issues. This one is highly relevant: pytorch/pytorch#14768 The suggested solution is, at the Linux command line, run Also, running To set permanently, edit |
Hmm, this problem as happened again :( I'm going to try once more... Before, I set |
This problem just happened again on I think the best solution to this problem is to remove PyTorch from the code #86
|
PR #307 removed the pytorch dataloader, so this issue is no longer relevant |
Describe the bug
prepare_ML_data.py
just crashed with this error:Interestingly, I had to hit
ctrl-c
to actually exit the script. Doing so threw this error, which I think tells us where in our code the crash occured:Versions:
To Reproduce
I fear this might be hard to reliably reproduce :)
The text was updated successfully, but these errors were encountered: