fix: use venv python for vLLM service startup

model-orchestrator/orchestrator_subprocess.py

@@ -102,11 +102,17 @@ async def start_model_process(model_name: str) -> bool:
         env.update({
             'HF_TOKEN': os.getenv('HF_TOKEN', ''),
             'PORT': str(port),
-            'HOST': '0.0.0.0'
+            'HOST': '0.0.0.0',
+            'MODEL_NAME': model_config.get('model_name', model_name)
         })
 
+        # Use venv python if it exists
+        script_dir = script_path.parent
+        venv_python = script_dir / 'venv' / 'bin' / 'python3'
+        python_cmd = str(venv_python) if venv_python.exists() else 'python3'
+
         proc = subprocess.Popen(
-            ['python3', str(script_path)],
+            [python_cmd, str(script_path)],
             env=env,
             stdout=subprocess.PIPE,
             stderr=subprocess.PIPE,