Update app.py

app.py

@@ -76,9 +76,7 @@ Set the environment variable `model` to change the model:
 ['AIML-TUDA/LlavaGuard-13B'](https://huggingface.co/AIML-TUDA/LlavaGuard-13B),
 ['AIML-TUDA/LlavaGuard-34B'](https://huggingface.co/AIML-TUDA/LlavaGuard-34B),
 """
-    # set_up_env_and_token(read=True)
     print(f"args: {gws.args}")
-    # set the huggingface login token
     controller_proc = start_controller()
     concurrency_count = int(os.getenv("concurrency_count", 5))
     api_key = os.getenv("token")
@@ -94,13 +92,15 @@ Set the environment variable `model` to change the model:
     models = [
         'LukasHug/LlavaGuard-7B-hf',
         'LukasHug/LlavaGuard-13B-hf',
-        'LukasHug/LlavaGuard-34B-hf',]
+        'LukasHug/LlavaGuard-34B-hf',
+        'liuhaotian/llava-v1.6-vicuna-7b',
+    ]
     bits = int(os.getenv("bits", 16))
-    model = os.getenv("model", models[0])
+    model = os.getenv("model", models[-1])
     available_devices = os.getenv("CUDA_VISIBLE_DEVICES", "0")
     model_path, model_name = model, model.split("/")[-1]
 
-    # worker_proc = start_worker(model_path, model_name, bits=bits)
+    worker_proc = start_worker(model_path, model_name, bits=bits)
 
 
     # Wait for worker and controller to start
@@ -122,7 +122,7 @@ Set the environment variable `model` to change the model:
         print(e)
         exit_status = 1
     finally:
-        # worker_proc.kill()
+        worker_proc.kill()
         controller_proc.kill()
 
         sys.exit(exit_status)