Skip to content

Use Float8TrainingOpConfig instead of removed FP8GroupedMMConfig alias #2573

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
tianyu-l merged 1 commit into from
Mar 14, 2026
Merged

Use Float8TrainingOpConfig instead of removed FP8GroupedMMConfig alias #2573

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
4 changes: 2 additions & 2 deletions torchtitan/components/quantization/float8.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -274,7 +274,7 @@ def convert(self, model: nn.Module):
from torchao.quantization.quant_api import quantize_

try:
from torchao.prototype.moe_training.config import FP8GroupedMMConfig
from torchao.prototype.moe_training.config import Float8TrainingOpConfig
except ImportError as e:
raise ImportError(
"torchao installation does not have MoE training support. Please install torchao nightly build."
Expand All @@ -293,7 +293,7 @@ def moe_module_filter_fn(mod: nn.Module, cur_fqn: str) -> bool:
model, ["_init_mean", "_init_std"], nn_module_cls=nn.Linear
)

config = FP8GroupedMMConfig()
config = Float8TrainingOpConfig()
quantize_(model, config=config, filter_fn=moe_module_filter_fn)

# Re-inject Linear protocol and re-attach attrs
Expand Down
Loading