Skip to content

[RFC] Seperate init_distributed_env from the Trainer.__init__ #2003

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
fegin merged 9 commits into from
Nov 8, 2025
Merged

[RFC] Seperate init_distributed_env from the Trainer.__init__ #2003

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
4d19355
Seperate init_distributed_env from the Trainer.__init__
fegin Nov 7, 2025
ea8ca6b
misc
fegin Nov 7, 2025
e1b7abe
misc
fegin Nov 7, 2025
05ff798
misc
fegin Nov 7, 2025
7e1bd6a
misc
fegin Nov 7, 2025
4cf92be
misc
fegin Nov 7, 2025
371f65f
linter
fegin Nov 7, 2025
cf11dd3
misc
fegin Nov 7, 2025
d97b7dc
misc
fegin Nov 8, 2025
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
16 changes: 14 additions & 2 deletions torchtitan/experiments/torchcomms/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,10 @@
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.

from torchtitan.distributed import ParallelDims
import os

from torchtitan.config import JobConfig
from torchtitan.distributed import ParallelDims, utils as dist_utils
from torchtitan.train import main, Trainer

from .parallel_dims import TorchCommsParallelDims
Expand All @@ -13,7 +16,16 @@
class TorchCommsTrainer(Trainer):
parallel_dims: TorchCommsParallelDims

def _create_parallel_dims(self, parallelism_config, world_size) -> ParallelDims:
def init_distributed_env(self, job_config: JobConfig) -> ParallelDims:
Copy link
Contributor

@tianyu-l tianyu-l Nov 7, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This PR doesn't do anything new with distributed backend. Are you saying we will modify this code for torchcomms?

Asking because I'm a bit concerned about sending whole job_config to this function, which seems a regression in terms of config usage hygiene.

dist_utils.init_distributed(
job_config.comm,
enable_cpu_backend=job_config.training.enable_cpu_offload,
base_folder=job_config.job.dump_folder,
)

world_size = int(os.environ["WORLD_SIZE"])
parallelism_config = job_config.parallelism

return TorchCommsParallelDims(
dp_shard=parallelism_config.data_parallel_shard_degree,
dp_replicate=parallelism_config.data_parallel_replicate_degree,
Expand Down
28 changes: 14 additions & 14 deletions torchtitan/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,20 +84,10 @@ def __init__(self, job_config: JobConfig):
# Device has to be set before creating TorchFT manager.
device_module.set_device(self.device)

# init distributed and build meshes
dist_utils.init_distributed(
job_config.comm,
enable_cpu_backend=job_config.training.enable_cpu_offload,
base_folder=job_config.job.dump_folder,
)

job_config.maybe_log()

world_size = int(os.environ["WORLD_SIZE"])
parallelism_config = job_config.parallelism
self.parallel_dims = parallel_dims = self._create_parallel_dims(
parallelism_config, world_size
)
# init distributed and build meshes
self.parallel_dims = parallel_dims = self.init_distributed_env(job_config)

world_mesh = parallel_dims.world_mesh
if parallel_dims.dp_enabled:
Expand Down Expand Up @@ -319,7 +309,8 @@ def __init__(self, job_config: JobConfig):
)

loss_parallel_enabled = (
parallel_dims.tp_enabled and not parallelism_config.disable_loss_parallel
parallel_dims.tp_enabled
and not job_config.parallelism.disable_loss_parallel
)
self.train_context = dist_utils.get_train_context(loss_parallel_enabled)
self.maybe_enable_amp = dist_utils.maybe_enable_amp(
Expand Down Expand Up @@ -367,7 +358,16 @@ def __init__(self, job_config: JobConfig):
f"(warmup {job_config.lr_scheduler.warmup_steps})"
)

def _create_parallel_dims(self, parallelism_config, world_size) -> ParallelDims:
def init_distributed_env(self, job_config: JobConfig) -> ParallelDims:
dist_utils.init_distributed(
job_config.comm,
enable_cpu_backend=job_config.training.enable_cpu_offload,
base_folder=job_config.job.dump_folder,
)

world_size = int(os.environ["WORLD_SIZE"])
parallelism_config = job_config.parallelism

return ParallelDims(
dp_shard=parallelism_config.data_parallel_shard_degree,
dp_replicate=parallelism_config.data_parallel_replicate_degree,
Expand Down
Loading