Skip to content

tinyfix: diffusion ci #13769

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
mickqian merged 4 commits into from
Nov 22, 2025
Merged

tinyfix: diffusion ci #13769

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6fd5a0f
Update test cases from ONE_GPU_CASES_A to B
yhyang201 Nov 22, 2025
69ff1bc
upd
mickqian Nov 22, 2025
05c1461
Merge remote-tracking branch 'origin/main' into tinyfix-diffusion-ci
mickqian Nov 22, 2025
3a177cd
upd baseline
mickqian Nov 22, 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
42 changes: 20 additions & 22 deletions python/sglang/multimodal_gen/test/server/perf_baselines.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,9 +5,9 @@
"description": "Reference numbers captured from the CI diffusion server baseline run"
},
"tolerances": {
"e2e": 0.05,
"e2e": 0.1,
"denoise_stage": 0.05,
"non_denoise_stage": 0.3,
"non_denoise_stage": 0.4,
"denoise_step": 0.2,
"denoise_agg": 0.08
},
Expand Down Expand Up @@ -41,7 +41,7 @@
"TimestepPreparationStage": 10.6,
"LatentPreparationStage": 11.8,
"DenoisingStage": 21202.6,
"DecodingStage": 512.6
"DecodingStage": 751.1
},
"denoise_step_ms": {
"0": 1077.77,
Expand Down Expand Up @@ -297,7 +297,6 @@
"expected_avg_denoise_ms": 1820.02,
"expected_median_denoise_ms": 1798.65
},

"wan2_2_ti2v_5b": {
"stages_ms": {
"InputValidationStage": 96.27,
Expand Down Expand Up @@ -387,26 +386,25 @@
},
"fast_hunyuan_video": {
"stages_ms": {
"InputValidationStage": 0.07,
"TextEncodingStage": 598.66,
"ConditioningStage": 0.01,
"TimestepPreparationStage": 57.26,
"LatentPreparationStage": 15.38,
"DenoisingStage": 3367.01,
"DecodingStage": 2457.57,
"per_frame_generation": null
"InputValidationStage": 0.09,
"TextEncodingStage": 845.64,
"ConditioningStage": 0.04,
"TimestepPreparationStage": 125.22,
"LatentPreparationStage": 29.34,
"DenoisingStage": 3860.64,
"DecodingStage": 2580.55
},
"denoise_step_ms": {
"0": 1577.86,
"1": 161.04,
"2": 406.81,
"3": 408.67,
"4": 403.55,
"5": 405.31
"0": 2063.08,
"1": 164.02,
"2": 406.99,
"3": 407.95,
"4": 407.51,
"5": 404.2
},
"expected_e2e_ms": 6521.03,
"expected_avg_denoise_ms": 560.54,
"expected_median_denoise_ms": 406.06
"expected_e2e_ms": 7487.87,
"expected_avg_denoise_ms": 642.29,
"expected_median_denoise_ms": 407.25
},
"wan2_2_i2v_a14b_2gpu": {
"stages_ms": {
Expand All @@ -415,7 +413,7 @@
"ConditioningStage": 0.02,
"TimestepPreparationStage": 2.2,
"LatentPreparationStage": 8.93,
"ImageVAEEncodingStage": 2075.47,
"ImageVAEEncodingStage": 2075.47,
"DenoisingStage": 382628.41,
"DecodingStage": 2820.89
},
Expand Down
4 changes: 2 additions & 2 deletions python/sglang/multimodal_gen/test/server/test_server_perf_b.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@
diffusion_server,
)
from sglang.multimodal_gen.test.server.testcase_configs import (
ONE_GPU_CASES_A,
ONE_GPU_CASES_B,
DiffusionTestCase,
)

Expand All @@ -25,7 +25,7 @@
class TestDiffusionPerformanceOneGpu(DiffusionPerformanceBase):
"""Performance tests for 1-GPU diffusion cases."""

@pytest.fixture(params=ONE_GPU_CASES_A, ids=lambda c: c.id)
@pytest.fixture(params=ONE_GPU_CASES_B, ids=lambda c: c.id)
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Nov 22, 2025

Choose a reason for hiding this comment

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

medium

While using ONE_GPU_CASES_B is correct here, the variable names ONE_GPU_CASES_A and ONE_GPU_CASES_B are not very descriptive. From testcase_configs.py, it appears 'A' is for image tests and 'B' is for video tests. Consider renaming them to ONE_GPU_IMAGE_CASES and ONE_GPU_VIDEO_CASES respectively. This would make the code more self-documenting and easier to maintain.

def case(self, request) -> DiffusionTestCase:
"""Provide a DiffusionTestCase for each 1-GPU test."""
return request.param
Loading