Skip to content

[Doc] Warning only in CUDA for lbfgs #74765

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 3 commits into from
Aug 25, 2025
Merged

[Doc] Warning only in CUDA for lbfgs #74765

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
229ec55
skip warning if not cuda device
HydrogenSulfate Aug 19, 2025
0390713
Merge branch 'PaddlePaddle:develop' into refine_lbfgs
HydrogenSulfate Aug 20, 2025
9f2ee0c
Merge branch 'PaddlePaddle:develop' into refine_lbfgs
HydrogenSulfate Aug 21, 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
7 changes: 5 additions & 2 deletions python/paddle/optimizer/lbfgs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,10 +59,13 @@ class _LbfgsStateDict(TypedDict):

def check_tf32_override():
"""Check and warn about TF32 acceleration status"""
if os.getenv("NVIDIA_TF32_OVERRIDE") != "0": # None or "1"
if (
paddle.device.is_compiled_with_cuda()
and os.getenv("NVIDIA_TF32_OVERRIDE") != "0"
): # None or "1"
warnings.warn(
"Warning! TF32 Tensor Cores are enabled by default on some NVIDIA GPUs for faster computation, "
"but may compromise numerical precision in specific cases, particularly with the L-BFGS optimizer. "
"but may compromise numerical precision in specific cases, particularly with the L-BFGS optimizer."
"To disable it, set: NVIDIA_TF32_OVERRIDE=0"
)

Expand Down