Skip to content

fix: refactor and fix Omni model tests to improve CI stability#1598

Merged
XuehaoSun merged 7 commits intomainfrom
xinhe/3-23b
Mar 25, 2026
Merged

fix: refactor and fix Omni model tests to improve CI stability#1598
XuehaoSun merged 7 commits intomainfrom
xinhe/3-23b

Conversation

@xin3he
Copy link
Copy Markdown
Contributor

@xin3he xin3he commented Mar 23, 2026

Description

  • Refactored Qwen2.5-Omni and Qwen3-Omni-MoE tests to use save_tiny_model and from_config for faster, checkpoint-free initialization.
  • Fixed issue where tests failed due to missing model assets or direct model object passing.
  • Improved log messages in BaseCompressor for unquantized layers.

Type of Change

  • Bug fix
  • New feature
  • Documentation update
  • Performance improvement
  • Code refactoring
  • Other (please specify):

Related Issues

Fixes or relates to #

Checklist Before Submitting

  • My code has been tested locally.
  • Documentation has been updated as needed.
  • New or updated tests are included where applicable.

Signed-off-by: Xin He <xin3.he@intel.com>
Copilot AI review requested due to automatic review settings March 23, 2026 07:38
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

This PR refactors CUDA integration tests for Omni models to rely on tiny, config-initialized models saved to disk (avoiding full checkpoint dependencies), and adjusts BaseCompressor logging around unquantized layers to be clearer.

Changes:

  • Refactored Qwen2.5-Omni and Qwen3-Omni-MoE CUDA tests to quantize from a saved tiny model path and reload via from_pretrained(...).
  • Extended test helpers to support from_config tiny-model initialization and generalized config layer/expert reduction.
  • Improved log messages for reporting unquantized layers during compression/quantization.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.

File Description
test/test_cuda/models/test_omni_model.py Switches Omni CUDA tests to quantize/reload from saved tiny model directories and uses tmp_path for outputs.
test/helpers.py Adds _reduce_config_layers and extends get_tiny_model/save_tiny_model with from_config and num_experts support.
test/fixtures.py Replaces Omni fixtures with session-scoped “tiny model path” fixtures built via save_tiny_model(..., from_config=True) and adds required assets.
auto_round/compressors/base.py Tweaks info-level logging to label “Unquantized layers” more explicitly.

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 23, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 23, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines will not run the associated pipelines, because the pull request was updated after the run command was issued. Review the pull request again and issue a new run command.

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 24, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

Signed-off-by: Xin He <xin3.he@intel.com>
@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 24, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 24, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 24, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 25, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@xin3he
Copy link
Copy Markdown
Contributor Author

xin3he commented Mar 25, 2026

/azp run Unit-Test-CUDA-AutoRound

@azure-pipelines
Copy link
Copy Markdown

Azure Pipelines successfully started running 1 pipeline(s).

@XuehaoSun XuehaoSun merged commit 8f10ac5 into main Mar 25, 2026
38 of 40 checks passed
@XuehaoSun XuehaoSun deleted the xinhe/3-23b branch March 25, 2026 04:10
xin3he added a commit that referenced this pull request Mar 25, 2026
Signed-off-by: Xin He <xin3.he@intel.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
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.

4 participants