Skip to content

port true bf16 training into forge experiment #1775

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
merged 1 commit into from
Sep 30, 2025
Merged

port true bf16 training into forge experiment #1775

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
7 changes: 6 additions & 1 deletion torchtitan/experiments/forge/engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,8 @@
import torchtitan.protocols.train_spec as train_spec_module
from torchtitan.components.checkpoint import CheckpointManager
from torchtitan.components.loss import rescale_accumulated_loss
from torchtitan.config import TORCH_DTYPE_MAP

from torchtitan.distributed import ParallelDims, utils as dist_utils
from torchtitan.protocols import BaseModelArgs
from torchtitan.tools import utils
Expand Down Expand Up @@ -114,7 +116,10 @@ def __init__(self, job_config: ForgeJobConfig):
# set the model args from training job configs
model_args.update_from_config(job_config)

with torch.device("meta"):
with (
torch.device("meta"),
utils.set_default_dtype(TORCH_DTYPE_MAP[job_config.training.dtype]),
):
model = self.train_spec.model_cls(model_args)

# calculate model size and flops per token
Expand Down