Fix qwen2_moe tests
#40494
Merged
Fix qwen2_moe tests
#40494
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ydshieh
commented
Aug 27, 2025
| out = model(input_ids).logits.float().cpu() | ||
| # Expected mean on dim = -1 | ||
| EXPECTED_MEAN = torch.tensor([[-4.2125, -3.6416, -4.9136, -4.3005, -4.9938, -3.4393, -3.5195, -4.1621]]) | ||
| EXPECTED_MEAN = torch.tensor([[-4.2106, -3.6411, -4.9111, -4.2840, -4.9950, -3.4438, -3.5262, -4.1624]]) |
Collaborator
Author
There was a problem hiding this comment.
changed because now we use fp16 (previously fp32)
Contributor
|
[For maintainers] Suggested jobs to run (before merge) run-slow: qwen2_moe |
ydshieh
commented
Aug 27, 2025
| def test_speculative_generation(self): | ||
| EXPECTED_TEXT_COMPLETION = ( | ||
| "To be or not to be, that is the question.\nThe answer is to be, of course. But what does it" | ||
| "To be or not to be, that is the question. Whether 'tis nobler in the mind to suffer the sl" |
Collaborator
Author
There was a problem hiding this comment.
the previous value never pass
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
zucchini-nlp
approved these changes
Aug 27, 2025
| assistant_model = model | ||
| assistant_model.generation_config.num_assistant_tokens = 2 | ||
| assistant_model.generation_config.num_assistant_tokens_schedule = "constant" | ||
| generated_ids = model.generate(input_ids, max_new_tokens=4, temperature=0) |
Member
There was a problem hiding this comment.
another test where assistant_model is not actually used 😄
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
The CI job of the tests in
Qwen2MoeIntegrationTestis killed from time to time due to the CPU memory limit (60 G), which is unclear to me why the usage may differ in different runs.This PR simply reuses the same model set once in the first test, except the
test_model_a2_7b_long_prompt_flash_attn, because it loads the model usingflash_attn.