Update app.py

app.py

@@ -24,6 +24,11 @@ def _get_args():
                         type=str,
                         default=DEFAULT_CKPT_PATH,
                         help='Checkpoint name or path, default to %(default)r')
+    parser.add_argument('-t',
+                        '--auth-token',
+                        type=str,
+                        default=AUTH_TOKEN,
+                        help='Authentication token for model repository, default to %(default)r')
     parser.add_argument('--cpu-only', action='store_true', help='Run demo with CPU only')
     parser.add_argument('--flash-attn2',
                         action='store_true',
@@ -54,12 +59,15 @@ def _load_model_processor(args):
     if args.flash_attn2:
         model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
             args.checkpoint_path,
-            use_auth_token=AUTH_TOKEN,
+            use_auth_token=args.auth_token,
             torch_dtype='auto',
             attn_implementation='flash_attention_2',
             device_map=device_map)
     else:
-        model = Qwen2_5_VLForConditionalGeneration.from_pretrained(args.checkpoint_path, device_map=device_map)
+        model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
+            args.checkpoint_path, 
+            use_auth_token=args.auth_token,
+            device_map=device_map)
 
     processor = AutoProcessor.from_pretrained(args.checkpoint_path)
     return model, processor