Skip to content

Enable FP8 DPA for MXFP8 recipe #4066

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
yaox12 merged 3 commits into from
Apr 10, 2026
+1 −1
Merged

Enable FP8 DPA for MXFP8 recipe #4066

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
9556e55
Enable FP8 DPA for MXFP8 recipe
vasunvidia Mar 20, 2026
d8ab252
Merge branch 'main' into vrengasamy/mxfp8Attn
vasunvidia Apr 6, 2026
83f33ab
fix lint
gautham-kollu Apr 7, 2026
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
2 changes: 1 addition & 1 deletion megatron/core/fp8_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -568,7 +568,7 @@ def get_fp8_recipe(config: TransformerConfig):
)
elif config.fp8_recipe == Fp8Recipe.mxfp8:
fp8_recipe = transformer_engine.common.recipe.MXFP8BlockScaling(
fp8_format=fp8_format
fp8_format=fp8_format, fp8_dpa=config.fp8_dot_product_attention
)
elif config.fp8_recipe == Fp8Recipe.custom:
assert config.fp8_quantizer_factory is not None
Expand Down
Loading