vllm/model_executor/models/llama4_eagle.py · f9df8b4ad77a933659c93cb1b923c8e09d76ea3a · OpenDAS / vllm_cscc

[Bugfix] Merge MM embeddings by index instead of token IDs (#16229) · 27d7638b

Cyrus Leung authored Sep 27, 2025


Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: NickLucche <nlucches@redhat.com>
Co-authored-by: Roger Wang <hey@rogerw.io>

27d7638b

llama4_eagle.py 8.92 KB

Replace llama4_eagle.py