Skip to content

(torchx/examples) set the correct number of gpus based on LOCAL_RANK or None (if cpu) #296

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

Closed
wants to merge 1 commit into from
Closed

(torchx/examples) set the correct number of gpus based on LOCAL_RANK or None (if cpu) #296

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
26 changes: 13 additions & 13 deletions torchx/examples/apps/lightning_classy_vision/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,22 +27,20 @@
import torch
from pytorch_lightning.callbacks import ModelCheckpoint
from pytorch_lightning.loggers import TensorBoardLogger

# ensure data and module are on the path
sys.path.append(".")

from torchx.examples.apps.lightning_classy_vision.data import (
TinyImageNetDataModule,
download_data,
create_random_data,
download_data,
)
from torchx.examples.apps.lightning_classy_vision.model import (
TinyImageNetModel,
export_inference_model,
)
from torchx.examples.apps.lightning_classy_vision.profiler import (
SimpleLoggingProfiler,
)
from torchx.examples.apps.lightning_classy_vision.profiler import SimpleLoggingProfiler


# ensure data and module are on the path
sys.path.append(".")


def parse_args(argv: List[str]) -> argparse.Namespace:
Expand Down Expand Up @@ -84,10 +82,6 @@ def parse_args(argv: List[str]) -> argparse.Namespace:
return parser.parse_args(argv)


def get_gpu_devices() -> int:
return torch.cuda.device_count()


def get_model_checkpoint(args: argparse.Namespace) -> Optional[ModelCheckpoint]:
if not args.output_path:
return None
Expand Down Expand Up @@ -138,10 +132,16 @@ def main(argv: List[str]) -> None:
# Initialize a trainer
num_nodes = int(os.environ.get("GROUP_WORLD_SIZE", 1))
num_processes = int(os.environ.get("LOCAL_WORLD_SIZE", 1))

if torch.cuda.is_available():
gpus = num_processes
else:
gpus = None

trainer = pl.Trainer(
num_nodes=num_nodes,
num_processes=num_processes,
gpus=get_gpu_devices(),
gpus=gpus,
accelerator="ddp",
logger=logger,
max_epochs=args.epochs,
Expand Down