fix: Remove vllm image cache

--bug=1052365 --user=刘瑞斌 【github#2353】vllm视觉模型修改最大tokens不生效 https://www.tapd.cn/57709429/s/1657667

Author: liuruibin

apps/setting/models_provider/impl/vllm_model_provider/model/image.py

@@ -18,3 +18,6 @@ def new_instance(model_type, model_name, model_credential: Dict[str, object], **
             stream_usage=True,
             **optional_params,
         )
+
+    def is_cache_model(self):
+        return False