vllm/model_executor/models/llama_eagle.py · 93c8672ceb06f6e9c282a96fcd85a7ce41293693 · OpenDAS / vllm_cscc · GitLab

Find file Blame History Permalink

[Bugfix] Fix spec decode memory regression after #28549 (#28819) · 93c8672c
Zhewen Li authored Nov 20, 2025
```
Signed-off-by: zhewenli <zhewenli@meta.com>
```
93c8672c

llama_eagle.py 7.75 KB