added auto-detection for the output type on the client side to be able to download binary files such as videos

source/model/PythonModel.py

@@ -77,6 +77,12 @@ class PythonModel(base.BaseModel):
             infer_api,
             methods=["POST"],
             tags=self.tags,
+            # summary=...,
+            # description=...,
+            response_class=fastapi.responses.StreamingResponse,
+            responses={
+                200: {"content": {self.output_type: {}}}
+            },
         )
 
     def _load(self) -> None: