benchmark mark accuracy use only 1 tensor

mindspore/lite/tools/benchmark/benchmark.cc

@@ -234,7 +234,12 @@ int Benchmark::CompareOutput() {
       MS_LOG(ERROR) << "Cannot find output node: " << nodeName.c_str() << " , compare output data fail.";
       return RET_ERROR;
     }
-    for (auto tensor : tensors) {
+    // make sure tensor size is 1
+    if (tensors.size() != 1) {
+      MS_LOG(ERROR) << "Only support 1 tensor with a name now.";
+      return RET_ERROR;
+    }
+    auto &tensor = tensors.front();
     MS_ASSERT(tensor->GetDataType() == DataType_DT_FLOAT);
     MS_ASSERT(tensor->GetData() != nullptr);
     float bias = CompareData(nodeName, tensor->shape(), static_cast<float *>(tensor->MutableData()));
@@ -246,7 +251,6 @@ int Benchmark::CompareOutput() {
       break;
     }
   }
-  }
 
   if (!hasError) {
     float meanBias;