diff --git a/paddle/fluid/operators/huber_loss_op.h b/paddle/fluid/operators/huber_loss_op.h
index fa21bd01cb052e5393385f0ef6c0203fc7c9e1a3..7000b5d3acc07d722c0129f86f0cfa2ee65b4288 100644
--- a/paddle/fluid/operators/huber_loss_op.h
+++ b/paddle/fluid/operators/huber_loss_op.h
@@ -41,7 +41,7 @@ struct HuberLossForward {
   T delta;
 };
 
-template <typename DeviceContext, typename T, typename AttrType = T>
+template <typename DeviceContext, typename T>
 class HuberLossKernel : public framework::OpKernel<T> {
  public:
   void Compute(const framework::ExecutionContext& context) const override {
@@ -49,7 +49,7 @@ class HuberLossKernel : public framework::OpKernel<T> {
     auto* in1 = context.Input<Tensor>("Y");
     auto* out0 = context.Output<Tensor>("Residual");
     auto* out1 = context.Output<Tensor>("Out");
-    auto delta = static_cast<T>(context.Attr<AttrType>("delta"));
+    auto delta = static_cast<T>(context.Attr<float>("delta"));
     auto& place =
         *context.template device_context<DeviceContext>().eigen_device();
 
@@ -86,7 +86,7 @@ struct HuberLossBackward {
   T delta;
 };
 
-template <typename DeviceContext, typename T, typename AttrType = T>
+template <typename DeviceContext, typename T>
 class HuberLossGradKernel : public framework::OpKernel<T> {
  public:
   void Compute(const framework::ExecutionContext& context) const override {
@@ -94,7 +94,7 @@ class HuberLossGradKernel : public framework::OpKernel<T> {
     auto* in1 = context.Input<Tensor>(framework::GradVarName("Out"));
     auto* out0 = context.Output<Tensor>(framework::GradVarName("X"));
     auto* out1 = context.Output<Tensor>(framework::GradVarName("Y"));
-    auto delta = static_cast<T>(context.op().Attr<AttrType>("delta"));
+    auto delta = static_cast<T>(context.op().Attr<float>("delta"));
     auto& place =
         *context.template device_context<DeviceContext>().eigen_device();