diff --git a/paddle/fluid/operators/sampling_id_op.cc b/paddle/fluid/operators/sampling_id_op.cc
index bf0739557bcda4a0b953f43df086b404b6fbe41b..128bc3f7fac5044c1c6503b1b7db9b76c5234b7b 100644
--- a/paddle/fluid/operators/sampling_id_op.cc
+++ b/paddle/fluid/operators/sampling_id_op.cc
@@ -24,16 +24,20 @@ class SamplingIdOp : public framework::OperatorWithKernel {
   using framework::OperatorWithKernel::OperatorWithKernel;
 
   void InferShape(framework::InferShapeContext* ctx) const override {
-    PADDLE_ENFORCE(ctx->HasInput("X"),
-                   "Input(X) of SamplingIdOp should not be null.");
-    PADDLE_ENFORCE(ctx->HasOutput("Out"),
-                   "Output(Out) of SamplingIdOp should not be null.");
-    PADDLE_ENFORCE_LT(ctx->Attrs().Get<float>("min"),
-                      ctx->Attrs().Get<float>("max"), "min must less then max");
+    OP_INOUT_CHECK(ctx->HasInput("X"), "Input", "X", "SampleIn");
+    OP_INOUT_CHECK(ctx->HasOutput("Out"), "Output", "X", "SampleOut");
+    PADDLE_ENFORCE_LT(
+        ctx->Attrs().Get<float>("min"), ctx->Attrs().Get<float>("max"),
+        platform::errors::InvalidArgument(
+            "min must less then max, but here min is %f, max is %f",
+            ctx->Attrs().Get<float>("min"), ctx->Attrs().Get<float>("max")));
 
     auto input_dims = ctx->GetInputDim("X");
-    PADDLE_ENFORCE(input_dims.size() == 2,
-                   "Input(X, Filter) should be 2-D tensor.");
+    PADDLE_ENFORCE_EQ(
+        input_dims.size(), 2,
+        platform::errors::InvalidArgument(
+            "Input(X, Filter) should be 2-D tensor. But X dim is %d",
+            input_dims.size()));
 
     auto dim0 = input_dims[0];
     framework::DDim dims = framework::make_ddim({dim0});
diff --git a/paddle/fluid/operators/sampling_id_op.h b/paddle/fluid/operators/sampling_id_op.h
index 133d3f72dbd6ab13c98d124369038309c94cba5b..5ec32c98f7f84abb255ec996d0cf6a58e6312ec3 100644
--- a/paddle/fluid/operators/sampling_id_op.h
+++ b/paddle/fluid/operators/sampling_id_op.h
@@ -36,9 +36,15 @@ class SamplingIdKernel : public framework::OpKernel<T> {
     const int batch_size = static_cast<int>(input->dims()[0]);
     const int width = static_cast<int>(input->dims()[1]);
 
-    PADDLE_ENFORCE_GE(batch_size, 0,
-                      "batch_size(dims[0]) must be nonnegative.");
-    PADDLE_ENFORCE_GE(width, 0, "width(dims[1]) must be nonnegative.");
+    PADDLE_ENFORCE_GE(
+        batch_size, 0,
+        platform::errors::InvalidArgument(
+            "batch_size(dims[0]) must be nonnegative. but it is %d.",
+            batch_size));
+    PADDLE_ENFORCE_GE(
+        width, 0,
+        platform::errors::InvalidArgument(
+            "width(dims[1]) must be nonnegative. but it is %d.", width));
 
     std::vector<T> ins_vector;
     framework::TensorToVector(*input, context.device_context(), &ins_vector);