Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

ImportError: cannot import name '_update_worker_pids'

See original GitHub issue

My setup:

Ubuntu 18.04.02
CUDA 10.0
PyTorch 1.1
Python3.5

Well, I tried to run the demo using: chmod +x demo_test.sh ./demo_test.sh

but the error came out:

Traceback (most recent call last): File “test.py”, line 11, in <module> from dataset import TestDataset File “/home/nhan/git_workspace/semantic-segmentation-pytorch/dataset.py”, line 4, in <module> import lib.utils.data as torchdata File “/home/nhan/git_workspace/semantic-segmentation-pytorch/lib/utils/data/init.py”, line 3, in <module> from .dataloader import DataLoader File “/home/nhan/git_workspace/semantic-segmentation-pytorch/lib/utils/data/dataloader.py”, line 3, in <module> from torch._C import _set_worker_signal_handlers, _update_worker_pids,
ImportError: cannot import name ‘_update_worker_pids’

What is the problem? Can you help me to solve it? Thanks.

Update 28 Mar 2019: Problem resolved. Rebuilding Pytorch v1.0.0 from source and it works now.

Issue Analytics

State:
Created 4 years ago
Reactions:2
Comments:5

Top GitHub Comments

37reactions

cplusxcommented, May 15, 2019

Though I see this issue has been solved by reinstalling pytorch version 1.0.1. But for those who has updated to 1.1.0 and does not want to downgrade. You can simply replace _update_worker_pids with _set_worker_pids. It works for me after replacement and I think it’s just a simple rename issue.

3reactions

AceCooooolcommented, Mar 27, 2019

This is due to update of dataloader in pytorch1.1（For example. this part is move to data._utils.worker.py file）, you need to rewrite this.