refactor(hpu_model_runner): restructure multimodal-related code #2066
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
wenbinc-Bin
reviewed
Oct 22, 2025
| [[1, image_h, int(mm_len / image_h)]]) | ||
| pixel_values = torch.randn( | ||
| image_grid_thw[0].prod(), | ||
| 1176) # TODO: figure out the variable name |
There was a problem hiding this comment.
we can find a way to change 1176 to "channel * temporal_patch_size * patch_size * patch_size" to avoid magic number.
|
|
||
| def compute_input_embedding(self, model, **kwargs): | ||
| input_ids = kwargs['input_ids'] | ||
| if model.config.model_type == 'qwen2_5_omni_thinker': |
There was a problem hiding this comment.
Because we refactor the code, we don't need to check it here. we can overwrite this function in qwen2_5_omni_thinker.py
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Essential Elements of an Effective PR Description Checklist
Purpose
The current multimodal model warmup logic is tightly coupled with HPUModelRunner, making it difficult to extend when adding new models.
Ideally, HPUModelRunner should only handle abstract warmup flow (e.g., wrap_multimodal_module_in_hpu(), init_multimodal_buckets(), create_dummy_inputs()) without any model-specific conditions.
This PR introduces an extensible MultimodalHandler class to unify the key steps of multimodal model warmup.
By subclassing MultimodalHandler and overriding specific methods, new multimodal models can be easily integrated.
The default implementation targets QwenVL-series models, and a separate Gemma3MultimodalHandler is provided to maintain backward compatibility with Gemma3.
Test Plan
Test Result