Skip to content

Fixing bug multi-gpu training #6299

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
glenn-jocher merged 2 commits into from
Jan 14, 2022
Merged

Fixing bug multi-gpu training #6299

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
eb6601b
Fixing bug multi-gpu training
Jan 14, 2022
9ee92ed
Update torch_utils.py for pep8
Jan 14, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion utils/torch_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,8 @@ def select_device(device='', batch_size=0, newline=True):
os.environ['CUDA_VISIBLE_DEVICES'] = '-1' # force torch.cuda.is_available() = False
elif device: # non-cpu device requested
assert torch.cuda.is_available(), 'CUDA unavailable' # check CUDA is available
assert torch.cuda.device_count() > int(device), f'invalid CUDA device {device} requested' # check index
device_list = [int(val) for val in device.replace(',', '')]
assert all([torch.cuda.device_count() > element for element in device_list]) , f'invalid CUDA device {device} requested' # check index
os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable (must be after asserts)

cuda = not cpu and torch.cuda.is_available()
Expand Down