-
Notifications
You must be signed in to change notification settings - Fork 101
multimodal model embedding fixes #759
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
libinta
commented
Dec 23, 2025
- remove scatter_mm_placeholders and gather_mm_placeholders from hpu_model_runner as upstream PR [Core][MM] Optimize encoder cache manager by operating with embeddings only vllm#30475 optimized it
- add HpuQwen3_VLForConditionalGeneration to use index_copy for _merge_multimodal_embeddings
Also overwrite qwen3_vl function to use _merge_multimodal_embeddings with index copy.
🚧 CI BlockedThe main CI workflow was not started for the following reason:
|
Format update
format
update for pre-commit error: vllm_gaudi/models/qwen3_vl.py:62:25: F821 Undefined name `_require_is_multimodal`
|
Please fix basic test |
I believe qwen3_vl only shows up from transformers 4.57.0 onwards. the CI tests seem to be using |
🚧 CI BlockedThe main CI workflow was not started for the following reason:
|
Signed-off-by: Libin Tang <litang@habana.ai>
🚧 CI BlockedThe main CI workflow was not started for the following reason:
|