Skip to content

[None][test] temporarily waive Cosmos3 B200 failures - #15195

Merged
bobboli merged 1 commit into
NVIDIA:mainfrom
bobboli:bli/waive-cosmos3-b200-failures
Jun 10, 2026
Merged

[None][test] temporarily waive Cosmos3 B200 failures#15195
bobboli merged 1 commit into
NVIDIA:mainfrom
bobboli:bli/waive-cosmos3-b200-failures

Conversation

@bobboli

@bobboli bobboli commented Jun 10, 2026

Copy link
Copy Markdown
Collaborator

Summary

  • Temporarily waive the eight DGX_B200 Cosmos3 failures from L0_MergeRequest_PR build 42291.
  • Keep this as a waive-only PR on top of latest main so the functional Cosmos3 config split fix can continue separately.

Test Plan

  • Commit pre-commit hooks passed, including waives duplicate and test-list validation.

Summary by CodeRabbit

  • Tests
    • Added temporary test waivers for specific Cosmos3 pipeline and transformer test cases pending configuration updates. Tests marked as skipped to allow continued development without blocking the build.

Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
@bobboli

bobboli commented Jun 10, 2026

Copy link
Copy Markdown
Collaborator Author

/bot skip --comment "Update waive"

@bobboli
bobboli enabled auto-merge (squash) June 10, 2026 03:34
@coderabbitai

coderabbitai Bot commented Jun 10, 2026

Copy link
Copy Markdown
Contributor

Review Change Stack

📝 Walkthrough

Walkthrough

This pull request adds eight temporary skip waiver entries to the test list for Cosmos3 visual-generation pipeline and transformer tests on DGX_B200 hardware. The additions mark specific test cases (FP8 load and transformer checkpoint/unit tests) as SKIP, each tagged with a comment indicating they are temporary waivers pending a Cosmos3 configuration split fix.

Changes

Cosmos3 Test Waivers

Layer / File(s) Summary
Cosmos3 Test Skip Entries
tests/integration/test_lists/waives.txt
Eight SKIP waiver lines added for full:DGX_B200/unittest/_torch/visual_gen/ Cosmos3 tests: TestCosmos3FP8Load::test_fp8_load_and_t2v, multiple TestCosmos3TransformerCheckpoint cases, and TestCosmos3Unit cases. All entries marked as temporary ToT main waivers pending Cosmos3 config split fix.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name Status Explanation Resolution
Description check ❓ Inconclusive The description addresses the core purpose and includes a test plan, but lacks structured sections and detailed reasoning required by the template. Provide a more structured description following the template: explain why the waive is necessary, link to the related issues/build failures, and clarify the relationship to the separate config split fix.
✅ Passed checks (4 passed)
Check name Status Explanation
Title check ✅ Passed The title clearly and specifically describes the main change: temporarily waiving Cosmos3 B200 test failures.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests

Comment @coderabbitai help to get the list of available commands and usage tips.

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Actionable comments posted: 1

🧹 Nitpick comments (1)
tests/integration/test_lists/waives.txt (1)

232-232: ⚡ Quick win

File-level skip entries likely ineffective.

Lines 232 and 234 skip entire test files without the ::TestClass::test_method notation. Searching the existing waives file shows no other bare-filename skips—all entries use the :: notation for specific tests. Since pytest generates nodeids like test_file.py::TestClass::test_method (confirmed in code snippet 3), these file-level entries likely won't match any collected items and are ineffective.

The specific test-level entries on lines 233 and 235–239 already cover the failing tests mentioned in the PR objectives. Recommend removing lines 232 and 234 unless there's evidence that file-level skips are expanded or matched by special logic.

Proposed change
-full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_pipeline.py SKIP (temporary ToT main waive; Cosmos3 config split fix pending)
 full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_pipeline.py::TestCosmos3FP8Load::test_fp8_load_and_t2v SKIP (temporary ToT main waive; Cosmos3 config split fix pending)
-full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_transformer.py SKIP (temporary ToT main waive; Cosmos3 config split fix pending)
 full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_transformer.py::TestCosmos3TransformerCheckpoint::test_load_fp8_quantization[FP8] SKIP (temporary ToT main waive; Cosmos3 config split fix pending)

Also applies to: 234-234

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@tests/integration/test_lists/waives.txt` at line 232, Remove the two
file-level waive entries that list
"full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_pipeline.py SKIP
(temporary ToT main waive; Cosmos3 config split fix pending)" (the bare-filename
skips) because pytest nodeids use the ::TestClass::test_method notation and the
existing specific test-level entries already cover the failing tests; delete
those bare file entries so only the specific test-level waives remain.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@tests/integration/test_lists/waives.txt`:
- Around line 232-239: The waive entries in
tests/integration/test_lists/waives.txt for Cosmos3 use inline reasons instead
of a tracking nvbugs link; create a single nvbugs tracking bug for the "Cosmos3
config split fix" and update all eight waive lines
(full:DGX_B200/unittest/_torch/visual_gen/test_cosmos33_pipeline.py and
full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_transformer.py entries
including specific tests like TestCosmos3FP8Load::test_fp8_load_and_t2v and
TestCosmos3TransformerCheckpoint::test_load_fp8_quantization[FP8] and
TestCosmos3Unit::test_model_structure/test_reset_cache/test_sanity_forward/test_sanity_forward_i2v_mask)
to append the short nvbugs URL format (e.g., SKIP (temporary ToT main waive;
https://nvbugs/XXXX)) replacing the inline text reason so all eight lines
reference the same nvbugs ID.

---

Nitpick comments:
In `@tests/integration/test_lists/waives.txt`:
- Line 232: Remove the two file-level waive entries that list
"full:DGX_B200/unittest/_torch/visual_gen/test_cosmos3_pipeline.py SKIP
(temporary ToT main waive; Cosmos3 config split fix pending)" (the bare-filename
skips) because pytest nodeids use the ::TestClass::test_method notation and the
existing specific test-level entries already cover the failing tests; delete
those bare file entries so only the specific test-level waives remain.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 01eb6e48-2d8f-4d80-a2bc-72a6b4a2d299

📥 Commits

Reviewing files that changed from the base of the PR and between 2763557 and 2c6e6ca.

📒 Files selected for processing (1)
  • tests/integration/test_lists/waives.txt

Comment thread tests/integration/test_lists/waives.txt
@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #53214 [ skip ] triggered by Bot. Commit: 2c6e6ca Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #53214 [ skip ] completed with state SUCCESS. Commit: 2c6e6ca
Skipping testing for commit 2c6e6ca

Link to invocation

@bobboli
bobboli merged commit 9c100bb into NVIDIA:main Jun 10, 2026
17 checks passed
This was referenced Aug 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants