fix: add --cuda-graph-max-bs to DSV3 FA3 FP8 KV cache test (#19307)

2026-02-25 10:27:31 -08:00
parent 0217e82a08
commit cc1ca61c81
1 changed files with 3 additions and 1 deletions
--- a/test/registered/mla/test_mla_deepseek_v3.py
+++ b/test/registered/mla/test_mla_deepseek_v3.py
@@ -110,7 +110,9 @@ class TestMLADeepseekV3Fa3Fp8Kvcache(CustomTestCase):
            "fp8_e4m3",
        ]
        if is_cuda():
-            other_args.extend(["--attention-backend", "fa3"])
+            other_args.extend(
+                ["--attention-backend", "fa3", "--cuda-graph-max-bs", "2"]
+            )
        cls.process = popen_launch_server(
            cls.model,
            cls.base_url,