[VLM] Support ViT Piecewise CUDA Graph for Qwen3-VL (#15320)

Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-12-20 21:00:07 +08:00
parent 1f1f05a85e
commit 019517a356
6 changed files with 235 additions and 66 deletions
@@ -19,6 +19,7 @@ from sglang.test.test_utils import (

 MODELS = [
    SimpleNamespace(model="Qwen/Qwen2.5-VL-7B-Instruct", mmmu_accuracy=0.60),
+    SimpleNamespace(model="Qwen/Qwen3-VL-8B-Instruct", mmmu_accuracy=0.60),
 ]