Update app.py

app.py

@@ -939,7 +939,6 @@ if __name__ == "__main__":
         attn_implementation='flash_attention_2',  # OR 'sdpa' for Ascend NPUs
         # torch_dtype=args.torch_dtype,
         # attn_implementation='sdpa',  # OR 'sdpa' for Ascend NPUs
-        use_auth_token=True,   
         low_cpu_mem_usage=True,
         trust_remote_code=True).eval().cuda()
     processor = AutoProcessor.from_pretrained(args.model_name, trust_remote_code=True)
@@ -948,7 +947,6 @@ if __name__ == "__main__":
     dualvitok = AutoModel.from_pretrained(args.tokenizer_path,
                                           torch_dtype=torch.float32,
                                           trust_remote_code=True,
-                                          use_auth_token=True,   
                                           ).eval().cuda()
     processor.set_vision_tokenizer(dualvitok)