Move mem_fraction_static adjustment for multimodal models to server_args.py & Fix session control & Other cleanups (#7748)

2025-07-04 16:33:33 -07:00
parent 975a5ec69c
commit 14229ccf8f
16 changed files with 339 additions and 137 deletions
--- a/test/srt/test_vision_openai_server_b.py
+++ b/test/srt/test_vision_openai_server_b.py
@@ -140,7 +140,7 @@ class TestGemma3itServer(TestOpenAIVisionServer):
            other_args=[
                "--trust-remote-code",
                "--mem-fraction-static",
-                "0.75",
+                "0.70",
                "--enable-multimodal",
            ],
        )