diff --git a/paddle/fluid/operators/concat_op.cc b/paddle/fluid/operators/concat_op.cc
index 68a52a79e4ce3..6095516f92fa5 100644
--- a/paddle/fluid/operators/concat_op.cc
+++ b/paddle/fluid/operators/concat_op.cc
@@ -244,4 +244,4 @@ REGISTER_OP_CPU_KERNEL(
     ops::ConcatGradKernel<paddle::platform::CPUDeviceContext,
                           paddle::platform::float16>,
     ops::ConcatGradKernel<paddle::platform::CPUDeviceContext, int>,
-    ops::ConcatKernel<paddle::platform::CPUDeviceContext, uint8_t>);
+    ops::ConcatGradKernel<paddle::platform::CPUDeviceContext, uint8_t>);
diff --git a/paddle/fluid/operators/concat_op.cu.cc b/paddle/fluid/operators/concat_op.cu.cc
index 8732556acb9fd..63025c3bd030f 100644
--- a/paddle/fluid/operators/concat_op.cu.cc
+++ b/paddle/fluid/operators/concat_op.cu.cc
@@ -33,4 +33,4 @@ REGISTER_OP_CUDA_KERNEL(
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, plat::float16>,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, int64_t>,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, int>,
-    ops::ConcatKernel<paddle::platform::CUDADeviceContext, uint8_t>);
+    ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, uint8_t>);