vllm/v1/worker/gpu_model_runner.py · 558f0907dc67c804c5821c92f4d64ed43d10489b · OpenDAS / vllm_cscc · GitLab

Find file Blame History Permalink

[attention][DCP] use AttentionImpl.need_to_return_lse_for_decode (#24372) · 558f0907
youkaichao authored Sep 07, 2025
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
558f0907

gpu_model_runner.py 160 KB