[Tests] [Bugfix] Skip MoE tests for models which cannot be imported - #2890
Conversation
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
|
👋 Hi! Thank you for contributing to llm-compressor. Please add the ready label when the PR is ready for review. Note: This is required to complete the testing suite, please only add the label once the PR is code complete and local testing has been performed. |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates tests/llmcompressor/modeling/test_linearize.py to skip test_linearize_moe instead of failing with an assertion when the model configuration or experts classes cannot be imported. The reviewer pointed out that the skip logic is missing for test_load_quantizable_moe, which was also supposed to be updated according to the PR description.
|
The quality checks have failed. Please run |
Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
|
The quality checks have failed. Please run |
dsikka
left a comment
dsikka
left a comment
There was a problem hiding this comment.
Wouldn’t it be easier to skip just based on version?
|
@dsikka I don't want to have to figure out which transformers version each of these model architectures were added |
|
Tests are failing because need vllm-project/compressed-tensors#767 or vllm-project/compressed-tensors#768 |
…2890) ## Purpose ## * Skip `test_linearize_moe` if model is not supported by local transformers version * Skip `test_load_quantizable_moe` if model is not supported by local transformers version * Reduce memory usage for `deepseek_v4` ## Testing ## * Tested that all tests pass, with 3 skipped when `transformers==5.9.0` ``` tests/llmcompressor/modeling/test_linearize.py ..s..............s....s........... [100%] ====================================== 31 passed, 3 skipped, 14 warnings in 1.79s ====================================== ``` --------- Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
3 participants
Purpose
test_linearize_moeif model is not supported by local transformers versiontest_load_quantizable_moeif model is not supported by local transformers versiondeepseek_v4Testing
transformers==5.9.0