Skip to content

Added single device training support #306

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Added single device training support #306

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e9193ae
Added single device training support
teng-li Feb 22, 2018
0c33fd9
Used --local_rank to specify the device to operate on
teng-li Mar 12, 2018
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
22 changes: 19 additions & 3 deletions imagenet/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,9 @@
help='url used to set up distributed training')
parser.add_argument('--dist-backend', default='gloo', type=str,
help='distributed backend')
parser.add_argument('--local_rank', default=-1, type=int,
help='If specified, the training will only run on the '
'single given GPU device of the local_rank')

best_prec1 = 0

Expand All @@ -63,12 +66,24 @@ def main():
global args, best_prec1
args = parser.parse_args()

if args.dist_url == "env://" and os.environ.get("WORLD_SIZE") is not None:
args.world_size = int(os.environ.get("WORLD_SIZE"))

args.distributed = args.world_size > 1

if args.distributed:
dist.init_process_group(backend=args.dist_backend, init_method=args.dist_url,
world_size=args.world_size)

if args.local_rank > -1 and args.local_rank < torch.cuda.device_count():
print("=> using single device: {} for training".format(args.local_rank))
dp_device_ids = [args.local_rank]
else:
dp_device_ids = None

if dp_device_ids:
torch.cuda.set_device(args.local_rank)

# create model
if args.pretrained:
print("=> using pre-trained model '{}'".format(args.arch))
Expand All @@ -79,13 +94,14 @@ def main():

if not args.distributed:
if args.arch.startswith('alexnet') or args.arch.startswith('vgg'):
model.features = torch.nn.DataParallel(model.features)
model.features = torch.nn.DataParallel(model.features,
device_ids=dp_device_ids)
model.cuda()
else:
model = torch.nn.DataParallel(model).cuda()
model = torch.nn.DataParallel(model, device_ids=dp_device_ids).cuda()
else:
model.cuda()
model = torch.nn.parallel.DistributedDataParallel(model)
model = torch.nn.parallel.DistributedDataParallel(model, device_ids=dp_device_ids)

# define loss function (criterion) and optimizer
criterion = nn.CrossEntropyLoss().cuda()
Expand Down