Skip to content

[VLM] Add Qwen3-VL generation test - #25185

Merged
DarkLight1337 merged 16 commits into
vllm-project:mainfrom
Isotr0py:qwen3-vl-test
Oct 29, 2025
Merged

DarkLight1337 merged 16 commits into
vllm-project:mainfrom
Isotr0py:qwen3-vl-test

Conversation

@Isotr0py

@Isotr0py Isotr0py commented Sep 18, 2025 •

Copy link
Copy Markdown
Member

Purpose

Test Plan

pytest -s -v tests/models/multimodal/generation/test_common.py -k qwen3_vl

Test Result

Test should pass


Essential Elements of an Effective PR Description Checklist
  • The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
  • The test plan, such as providing test command.
  • The test results, such as pasting the results comparison before and after, or e2e results
  • (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
  • (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
@mergify mergify Bot added multi-modality Related to multi-modality (#4194) qwen Related to Qwen models labels Sep 18, 2025
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
@Isotr0py
Isotr0py marked this pull request as ready for review September 19, 2025 06:22
@DarkLight1337

Copy link
Copy Markdown
Member

Should we wait until model release before actually merging this?

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
@Isotr0py

Copy link
Copy Markdown
Member Author

Should we wait until model release before actually merging this?

I'm fine to wait until model release and Transformers upgrade, because this test won't run on CI without them.

The goal of PR is just to provide a test for generation correctness validation, so that we can run correctness test for other Qwen3-VL related PRs locally to avoid regression. So whether merge this PR or not won't be a big deal IMO.

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
@DarkLight1337

Copy link
Copy Markdown
Member

Can you update this PR and get the tests passing so we can run the test locally to check further modifications to the model? This would help avoid issues such as #27705 cc @ywang96

Signed-off-by: Roger Wang <hey@rogerw.io>
@ywang96

ywang96 commented Oct 29, 2025

Copy link
Copy Markdown
Member

I'll work on this PR since @Isotr0py might be busy

Signed-off-by: Roger Wang <hey@rogerw.io>
Signed-off-by: Roger Wang <hey@rogerw.io>
@ywang96

ywang96 commented Oct 29, 2025

Copy link
Copy Markdown
Member

Test results from pytest -s -v tests/models/multimodal/generation/test_common.py -k qwen3_vl

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
================================================== 12 passed, 327 deselected, 28 warnings in 581.95s (0:09:41) ===================================================

Signed-off-by: Roger Wang <hey@rogerw.io>
Signed-off-by: Roger Wang <hey@rogerw.io>
@ywang96 ywang96 added the ready ONLY add when PR is ready to merge/full CI is needed label Oct 29, 2025
Signed-off-by: Roger Wang <hey@rogerw.io>
@DarkLight1337
DarkLight1337 enabled auto-merge (squash) October 29, 2025 10:03
@DarkLight1337
DarkLight1337 merged commit ad3ec89 into vllm-project:main Oct 29, 2025
42 checks passed
@Isotr0py
Isotr0py deleted the qwen3-vl-test branch October 29, 2025 13:02
ilmarkov pushed a commit to neuralmagic/vllm that referenced this pull request Nov 7, 2025
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
ZhengHongming888 pushed a commit to ZhengHongming888/vllm that referenced this pull request Nov 8, 2025
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
rtourgeman pushed a commit to rtourgeman/vllm that referenced this pull request Nov 10, 2025
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
devpatelio pushed a commit to SumanthRH/vllm that referenced this pull request Nov 29, 2025
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
mystous pushed a commit to mystous/vllm_hybrid that referenced this pull request May 10, 2026
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
my-other-github-account pushed a commit to my-other-github-account/vllm that referenced this pull request May 15, 2026
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
0826joyce pushed a commit to 0826joyce/vllm-serving-optimization that referenced this pull request May 19, 2026
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
plasticchris pushed a commit to plasticchris/vllm that referenced this pull request Jul 20, 2026
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

multi-modality Related to multi-modality (#4194) qwen Related to Qwen models ready ONLY add when PR is ready to merge/full CI is needed

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants