switch to Qwen3-VL-8B-Instruct, fix start.sh arg, enable tool calling with hermes parser

server.py

@@ -34,18 +34,18 @@ if __name__ == "__main__":
     os.environ["CUDA_VISIBLE_DEVICES"] = gpu
     os.environ["FLASHINFER_DISABLE_VERSION_CHECK"] = "1"
 
-    model = str(Path(__file__).resolve().parent / "models/Qwen2.5-7B-Instruct")
+    model = str(Path(__file__).resolve().parent / "models/Qwen3-VL-8B-Instruct")
 
     parser = make_arg_parser(FlexibleArgumentParser())
     args = parser.parse_args([
         "--model", model,
         "--host", "0.0.0.0",
         "--port", "8000",
-        "--served-model-name", "Qwen2.5-7B-Instruct",
+        "--served-model-name", "Qwen3-VL-8B-Instruct",
         "--trust-remote-code",
         "--max-model-len", "32768",
         "--dtype", "bfloat16",
-        "--gpu-memory-utilization", "0.85",
+        "--gpu-memory-utilization", "0.70",
         "--enforce-eager",
         "--enable-auto-tool-choice",
         "--tool-call-parser", "hermes",