fix(e2e): re-point HunyuanImage3 offline test to text_to_image.py

[chethanuk]

Purpose

Fixes nightly/CI failure tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py::test_hunyuan_image3_pixel_accuracy_offline — can't open file '.../hunyuan_image3/end2end.py': [Errno 2].

Root cause: commit a7ab91fb8 (vllm-project#5559) deleted examples/offline_inference/hunyuan_image3/end2end.py and migrated HunyuanImage-3.0 to the shared examples/offline_inference/text_to_image/text_to_image.py, but tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py:76 still invoked the deleted path, exiting with status 2 (/usr/bin/python3: can't open file ...).

Fix (one file, tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py):

• _OFFLINE_SCRIPT re-pointed to examples/offline_inference/text_to_image/text_to_image.py.
• _run_vllm_omni_hunyuan_image3_offline flag remap: --modality dropped, --prompts → --prompt, --steps → --num-inference-steps, --bot-task none + --sys-type en_unified → --extra-body '{"bot_task": null}' + --use-system-prompt en_unified + --trust-remote-code, and --output now a direct file path (no output_*.png glob/re-save). See text_to_image.py:106,128,140,323,364,383.

Fixes vllm-project#7495

Test Plan

vLLM Version: n/a

vLLM-Omni Commit: 0a857ad (this branch, 1 commit vs main)

Local (non-GPU):

• python3 -m py_compile tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py — OK
• git grep -n "hunyuan_image3.*end2end" — only tests/buildkite/test_vllm_omni_package_discovery.py:59 (end2end_probe.py fixture) remains
• Flag existence check: --prompt :106, --output :128, --num-inference-steps :140, --extra-body :323, --use-system-prompt :364 (en_unified in choices), --trust-remote-code :383 in text_to_image.py

GPU (exact CI commands, per plan):

• CUDA 4×H100 (.buildkite/cuda/test-nightly.yml:286): pytest -s -v tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py -m full_model --run-level full_model (add -k offline for local)
• NPU 4×A3 (.buildkite/npu/test-npu-nightly.yml:179): -m "full_model and A3" --run-level full_model

Note: old script let pipeline_hunyuan_image3.py:1235 build the generator, new script seeds via torch.Generator(device).manual_seed(seed) (text_to_image.py:489); seeding triage: if offline SSIM/PSNR regresses while ..._online passes, check noise seeding first before touching thresholds.

Test Result

• py_compile OK, worktree/primary clean, single-file diff (10+, 14-).
• No runtime on 4×H100/A3 in this env; full pytest -m full_model to be confirmed on CI.

BEFORE SUBMITTING: read CONTRIBUTING.md and run the precheck-pr skill with the code agent for a self-check against project conventions.

[codeant-ai]

🤖 CodeAnt AI — Review Status

Status	Commit	Started (UTC)	Finished (UTC)
✅ Reviewed your PR	0a857ad	Sep 14, 2026 · 14:27	14:31

[codeant-ai]

---

Thanks for using CodeAnt! 🎉

We're free for open-source projects. if you're enjoying it, help us grow by sharing.

Share on X ·
Reddit ·
LinkedIn

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Advanced

Run ID: 47ba8658-bac8-43d6-8e23-91a17dd6ba42

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Advanced

Run ID: 68631ebd-ef48-458c-a2ce-775bf06be22f

📥 Commits

Reviewing files that changed from the base of the PR and between 58adeec and 0a857ad.

📒 Files selected for processing (1)

• tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py

Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review.

---

📝 Walkthrough

Walkthrough

The HunyuanImage-3 offline pixel-accuracy test now invokes the generic text-to-image example with updated CLI options. It passes system-prompt and request-body values, enables remote code, and validates the image at the specified output path.

Changes

HunyuanImage-3 accuracy test

Layer / File(s)	Summary
Generic text-to-image runner integration tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py	The test selects the generic offline text-to-image script and uses --prompt and --num-inference-steps.
Request options and output validation tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py	The test replaces the previous bot-task and system-type flags, adds --trust-remote-code, and loads the image from the exact output path after checking that the path exists.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Suggested reviewers: yenuo26

Merge Risk: ⚪ Minimal · up to 0a857

The accuracy test now uses the supported shared runner interface and reads the exact image path it produces.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 2 functions across 1 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the HunyuanImage-3 offline test fix and specifies that it now uses text_to_image.py.
Description check	✅ Passed	The description is directly related to the changeset. It explains the deleted script failure, the replacement script, the CLI updates, and the planned validation.
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.

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codeant-ai]

User description

Purpose

Fixes nightly/CI failure tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py::test_hunyuan_image3_pixel_accuracy_offline — can't open file '.../hunyuan_image3/end2end.py': [Errno 2].

Root cause: commit a7ab91fb8 (vllm-project#5559) deleted examples/offline_inference/hunyuan_image3/end2end.py and migrated HunyuanImage-3.0 to the shared examples/offline_inference/text_to_image/text_to_image.py, but tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py:76 still invoked the deleted path, exiting with status 2 (/usr/bin/python3: can't open file ...).

Fix (one file, tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py):

• _OFFLINE_SCRIPT re-pointed to examples/offline_inference/text_to_image/text_to_image.py.
• _run_vllm_omni_hunyuan_image3_offline flag remap: --modality dropped, --prompts → --prompt, --steps → --num-inference-steps, --bot-task none + --sys-type en_unified → --extra-body '{"bot_task": null}' + --use-system-prompt en_unified + --trust-remote-code, and --output now a direct file path (no output_*.png glob/re-save). See text_to_image.py:106,128,140,323,364,383.

Fixes vllm-project#7495

Test Plan

vLLM Version: n/a

vLLM-Omni Commit: 0a857ad (this branch, 1 commit vs main)

Local (non-GPU):

• python3 -m py_compile tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py — OK
• git grep -n "hunyuan_image3.*end2end" — only tests/buildkite/test_vllm_omni_package_discovery.py:59 (end2end_probe.py fixture) remains
• Flag existence check: --prompt :106, --output :128, --num-inference-steps :140, --extra-body :323, --use-system-prompt :364 (en_unified in choices), --trust-remote-code :383 in text_to_image.py

GPU (exact CI commands, per plan):

• CUDA 4×H100 (.buildkite/cuda/test-nightly.yml:286): pytest -s -v tests/e2e/accuracy/test_hunyuan_image3_pixel_accuracy.py -m full_model --run-level full_model (add -k offline for local)
• NPU 4×A3 (.buildkite/npu/test-npu-nightly.yml:179): -m "full_model and A3" --run-level full_model

Note: old script let pipeline_hunyuan_image3.py:1235 build the generator, new script seeds via torch.Generator(device).manual_seed(seed) (text_to_image.py:489); seeding triage: if offline SSIM/PSNR regresses while ..._online passes, check noise seeding first before touching thresholds.

Test Result

• py_compile OK, worktree/primary clean, single-file diff (10+, 14-).
• No runtime on 4×H100/A3 in this env; full pytest -m full_model to be confirmed on CI.

BEFORE SUBMITTING: read CONTRIBUTING.md and run the precheck-pr skill with the code agent for a self-check against project conventions.

---

CodeAnt-AI Description

Restore HunyuanImage-3 offline pixel-accuracy testing

What Changed

• The offline accuracy test now uses the shared text-to-image example instead of the removed HunyuanImage-3 script.
• Updated generation options to match the shared example, including prompt, inference-step, system-prompt, and model configuration.
• Generated images are saved to and verified at the expected file path, allowing the test to compare them with the baseline image.

Impact

✅ Fewer nightly CI failures
✅ Restored HunyuanImage-3 offline accuracy coverage
✅ Reliable generated-image validation

💡 Usage Guide

Checking Your Pull Request

Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.

Talking to CodeAnt AI

Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:

@codeant-ai ask: Your question here

This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.

Example

@codeant-ai ask: Can you suggest a safer alternative to storing this secret?

Preserve Org Learnings with CodeAnt

You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:

@codeant-ai: Your feedback here

This helps CodeAnt AI learn and adapt to your team's coding style and standards.

Example

@codeant-ai: Do not flag unused imports.

Retrigger review

Ask CodeAnt AI to review the PR again, by typing:

@codeant-ai: review

Check Your Repository Health

To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.

[codeant-ai]

Skipping CodeAnt AI review — this PR changes more than 100 files, which usually means a migration, codemod, or vendored drop. Line-level review on diffs this large produces duplicate findings on the same rewrite pattern and drowns out anything that actually matters.

If you still want a review, comment @codeant-ai : review. For better signal, consider splitting the PR into smaller chunks.