Skip to content

[ROCm] Relax tolerances for FP8 unit test for fp16 and bf16 cases #7655

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

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

[ROCm] Relax tolerances for FP8 unit test for fp16 and bf16 cases #7655

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 tests/unit/runtime/half_precision/test_fp8.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,8 +85,8 @@ def run_zero(stage, model_dtype):
rtol, atol = 1e-07, 1e-05

# Relax tolerance only for ROCm + FP16
if is_rocm_pytorch() and model_dtype == torch.float16:
rtol, atol = 3e-07, 3e-05
if is_rocm_pytorch() and base_datatype in ["fp16", "bf16"]:
rtol, atol = 1e-07, 1e-04
Copy link
Contributor

@jithunnair-amd jithunnair-amd Oct 30, 2025

Choose a reason for hiding this comment

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

@rraminen Suggest figuring out if this is HW-specific e.g. MI200/MI300?


# config
zero_stage = [0, 1, 2, 3]
Expand Down