autoconvert tensor rt engine from onnx to specific CUDA gpu

Azaion.Inference/remote_command_handler.pyx

@@ -76,7 +76,7 @@ cdef class RemoteCommandHandler:
         with self._context.socket(zmq.DEALER) as socket:
             socket.connect("inproc://backend")
             socket.send_multipart([client_id, data])
-            constants.log(<str>f'Sent {len(data)} bytes.', client_id)
+            # constants.log(<str>f'Sent {len(data)} bytes.', client_id)
 
     cdef stop(self):
         self._shutdown_event.set()