diff --git a/demo-client/src/bert_service.cpp b/demo-client/src/bert_service.cpp
index 11a5276fd3ff653f8a2d6127bdcc471e4ba5b07b..f272078698e9921c0fad0b65666589d4c8c6a2e6 100644
--- a/demo-client/src/bert_service.cpp
+++ b/demo-client/src/bert_service.cpp
@@ -94,6 +94,7 @@ int create_req(Request* req,
       }
     }
     ins->set_max_seq_len(max_seq_len);
+    ins->set_emb_size(emb_size);
   }
   return 0;
 }
diff --git a/demo-serving/op/bert_service_op.cpp b/demo-serving/op/bert_service_op.cpp
index f4713782bf54e894efdf391c05231d77accf7019..f34d113a59fc79a7badaaf75f3b839e3200c89f6 100644
--- a/demo-serving/op/bert_service_op.cpp
+++ b/demo-serving/op/bert_service_op.cpp
@@ -127,13 +127,13 @@ int BertServiceOp::inference() {
 #if 0  // print request
   std::ostringstream oss;
   for (int j = 0; j < 3; j++) {
-    int64_t* example = reinterpret_cast<int64_t*>(*in)[j].data.data();
+    int64_t* example = reinterpret_cast<int64_t*>((*in)[j].data.data());
     for (uint32_t i = 0; i < MAX_SEQ_LEN; i++) {
         oss << *(example + i) << " ";
     }
     oss << ";";
   }
-  float* example = reinterpret_cast<float*>(*in)[3].data.data();
+  float* example = reinterpret_cast<float*>((*in)[3].data.data());
   for (int i = 0; i < MAX_SEQ_LEN; i++) {
     oss << *(example + i) << " ";
   }
diff --git a/sdk-cpp/proto/bert_service.proto b/sdk-cpp/proto/bert_service.proto
index fbf07e705985a336333984622a00007667d67dd3..01168560e1d1fb698e326a966aeb553decbf538c 100644
--- a/sdk-cpp/proto/bert_service.proto
+++ b/sdk-cpp/proto/bert_service.proto
@@ -38,7 +38,7 @@ message BertResInstance { repeated Embedding_values instances = 1; };
 message Response {
   repeated BertResInstance instances = 1;
   optional int64 op_time = 2;
-  optional int64 infert_time = 3;
+  optional int64 infer_time = 3;
 };
 
 service BertService {