Skip to content
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

Jump to bottom

[train] fix transformers example for multi-gpu #24832

Merged
merged 5 commits into from Jun 9, 2022
Merged

[train] fix transformers example for multi-gpu #24832

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
47c3d2b
[train] fix transformers example for multi-gpu
matthewdeng May 16, 2022
b538bd0
Merge branch 'master' into pr/matthewdeng/24832
Yard1 Jun 8, 2022
9405aa1
Use GPU for transformers example
Yard1 Jun 8, 2022
98db3cf
Fix
Yard1 Jun 8, 2022
db798f4
Fix for CPU
Yard1 Jun 8, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
14 changes: 13 additions & 1 deletion python/ray/train/BUILD
View file
Open in desktop
Expand Up @@ -49,7 +49,19 @@ py_test(
)

py_test(
name = "transformers_example",
name = "transformers_example_gpu",
size = "large",
main = "examples/transformers/transformers_example.py",
srcs = ["examples/transformers/transformers_example.py"],
tags = ["team:ml", "exclusive", "tune", "gpu_only"],
deps = [":train_lib"],
args = ["--model_name_or_path=bert-base-cased", "--task_name=mrpc",
"--max_length=32", "--per_device_train_batch_size=64",
"--max_train_steps=2", "--start_local", "--num_workers=2", "--use_gpu"]
)

py_test(
name = "transformers_example_cpu",
size = "large",
main = "examples/transformers/transformers_example.py",
srcs = ["examples/transformers/transformers_example.py"],
Expand Down
2 changes: 2 additions & 0 deletions python/ray/train/examples/transformers/cluster.yaml
View file
Open in desktop
Expand Up @@ -19,6 +19,7 @@ available_node_types:
ray.head.default:
min_workers: 0
max_workers: 0
resources: {}
node_config:
InstanceType: g3.8xlarge
ImageId: latest_dlami
Expand All @@ -33,6 +34,7 @@ available_node_types:
ray.worker.default:
min_workers: 3
max_workers: 3
resources: {}
node_config:
InstanceType: g3.8xlarge
ImageId: latest_dlami
Expand Down
6 changes: 5 additions & 1 deletion python/ray/train/examples/transformers/transformers_example.py
View file
Open in desktop
Expand Up @@ -228,10 +228,14 @@ def parse_args():


def train_func(config: Dict[str, Any]):
# Accelerator reads from this environment variable for GPU placement.
os.environ["LOCAL_RANK"] = str(ray.train.local_rank())
os.environ["WORLD_SIZE"] = str(ray.train.world_size())

args = config["args"]
# Initialize the accelerator. We will let the accelerator handle device
# placement for us in this example.
accelerator = Accelerator()
accelerator = Accelerator(cpu=not args.use_gpu)
# Make one log on every process with the configuration for debugging.
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
Expand Down