diff --git a/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/batch_norm_gpu_kernel.h b/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/batch_norm_gpu_kernel.h
index 02ff79cc7e..8e499ff3dc 100644
--- a/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/batch_norm_gpu_kernel.h
+++ b/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/batch_norm_gpu_kernel.h
@@ -96,6 +96,7 @@ class BatchNormGpuKernel : public GpuKernel {
     }
 
     InitResource();
+    is_train_ = GetAttr<bool>(kernel_node, "is_training");
     if (is_train_) {
       mode_ = CUDNN_BATCHNORM_SPATIAL_PERSISTENT;
     } else {
@@ -133,7 +134,6 @@ class BatchNormGpuKernel : public GpuKernel {
     }
     SetTensorDescriptor(format, shape);
     InitSizeLists();
-    is_train_ = GetAttr<bool>(kernel_node, "is_training");
     return true;
   }
 
@@ -229,8 +229,8 @@ class BatchNormGpuKernel : public GpuKernel {
     }
 
     output_size_list_.push_back(output_size_);   // output
-    output_size_list_.push_back(reserve_size_);  // reserve space
     output_size_list_.push_back(para_size_);     // save scale
+    output_size_list_.push_back(reserve_size_);  // reserve space
     output_size_list_.push_back(para_size_);     // save mean
     output_size_list_.push_back(para_size_);     // save variance