[Test] Add voice or language test case for Qwen3-omni and Qwen-tts#1844
[Test] Add voice or language test case for Qwen3-omni and Qwen-tts#1844gcanlin merged 16 commits intovllm-project:mainfrom
Conversation
- Added OpenCC dependency for text conversion in preprocess_text function. - Introduced functions to estimate voice gender from audio and validate against expected gender in assert_omni_response. - Refactored audio merging logic into a separate function for clarity. - Updated test cases to include voice gender specifications and improved parameter handling for model configurations. Signed-off-by: yenuo26 <410167048@qq.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 16e54f14bd
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| if f0 > 170.0: | ||
| half = f0 / 2.0 | ||
| if 75.0 <= half <= 185.0: | ||
| f0 = half |
There was a problem hiding this comment.
Avoid unconditional octave halving in gender estimator
This octave-correction branch halves any f0 > 170 when the halved value is in 75..185, which incorrectly maps most genuine female pitches (typically ~180–255 Hz) into the male range before classification. In practice, a correctly estimated female voice around 200–230 Hz gets forced to ~100–115 Hz and then labeled male, so the new female-voice validations can fail even when the model output is correct.
Useful? React with 👍 / 👎.
Signed-off-by: wangyu <410167048@qq.com>
…and Qwen3-omni Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
|
@congw729 please help to add nightly-test label |
…gging and efficiency. Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <53896905+yenuo26@users.noreply.github.com>
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
|
success in nightly-test: |
Signed-off-by: wangyu <410167048@qq.com>
Signed-off-by: wangyu <410167048@qq.com>
…llm-project#1844) Signed-off-by: yenuo26 <410167048@qq.com> Signed-off-by: wangyu <410167048@qq.com> Signed-off-by: wangyu <53896905+yenuo26@users.noreply.github.com>


PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED.
Purpose
1.Add speaker and language test cases for Qwen3-omni
2.Add voice case-sensitive test cases for Qwen-tts
Test Plan
2.Qwen-tts
Test Result
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model. Please runmkdocs serveto sync the documentation editions to./docs.BEFORE SUBMITTING, PLEASE READ https://github.com/vllm-project/vllm-omni/blob/main/CONTRIBUTING.md (anything written below this line will be removed by GitHub Actions)