add double grad for square op (#17173)

* add double grad for square. test=develop

* formax code. test=develop

* fix for grad sum. test=develop

* refine shape. test=develop

* refine extract. test=develop

paddle/fluid/operators/activation_op.cc

@@ -597,11 +597,23 @@ REGISTER_ACTIVATION_OP_MAKER(Square, SquareDoc);
 REGISTER_ACTIVATION_OP_MAKER(Softplus, SoftplusDoc);
 REGISTER_ACTIVATION_OP_MAKER(Softsign, SoftsignDoc);
 
+template <ActBwdOpFwdDeps kDepValue>
 class ActivationOpDoubleGrad : public framework::OperatorWithKernel {
  public:
   using framework::OperatorWithKernel::OperatorWithKernel;
 
   void InferShape(framework::InferShapeContext* ctx) const override {
+    if (static_cast<int>(kDepValue) & static_cast<int>(kDepX)) {
+      if (ctx->HasOutput("DX")) {
+        ctx->ShareDim("X", "DX");
+        ctx->ShareLoD("X", "DX");
+      }
+      if (ctx->HasOutput("DDOut")) {
+        ctx->ShareDim("X", "DDOut");
+        ctx->ShareLoD("X", "DDOut");
+      }
+    }
+    if (static_cast<int>(kDepValue) & static_cast<int>(kDepOut)) {
       if (ctx->HasOutput("DOut")) {
         ctx->ShareDim("Out", "DOut");
         ctx->ShareLoD("Out", "DOut");
@@ -611,27 +623,6 @@ class ActivationOpDoubleGrad : public framework::OperatorWithKernel {
         ctx->ShareLoD("Out", "DDOut");
       }
     }
-
- protected:
-  framework::OpKernelType GetExpectedKernelType(
-      const framework::ExecutionContext& ctx) const override {
-    return GetKernelType(ctx, *this, "Out");
-  }
-};
-
-class LeakyReluDoubleGrad : public framework::OperatorWithKernel {
- public:
-  using framework::OperatorWithKernel::OperatorWithKernel;
-
-  void InferShape(framework::InferShapeContext* ctx) const override {
-    if (ctx->HasOutput("DX")) {
-      ctx->ShareDim("X", "DX");
-      ctx->ShareLoD("X", "DX");
-    }
-    if (ctx->HasOutput("DDOut")) {
-      ctx->ShareDim("X", "DDOut");
-      ctx->ShareLoD("X", "DDOut");
-    }
   }
 
  protected:
@@ -690,6 +681,33 @@ class LeakyReluDoubleGradMaker
   }
 };
 
+// square Grad: dx=2x*dy
+// square GradGrad: ddy=2x*ddx, dx=2dy*ddx
+class SquareDoubleGradMaker
+    : public ::paddle::framework::SingleGradOpDescMaker {
+ public:
+  using ::paddle::framework::SingleGradOpDescMaker::SingleGradOpDescMaker;
+
+ protected:
+  std::unique_ptr<::paddle::framework::OpDesc> Apply() const override {
+    auto* op = new ::paddle::framework::OpDesc();
+    op->SetType("square_grad_grad");
+    op->SetInput("X", Input("X"));
+    // Out@GRAD: dy
+    op->SetInput("DOut", Input(framework::GradVarName("Out")));
+    // X@GRAD@GRAD: ddx
+    op->SetInput("DDX", OutputGrad(framework::GradVarName("X")));
+
+    op->SetAttrMap(Attrs());
+
+    // X@GRAD: dx
+    op->SetOutput("DX", InputGrad("X"));
+    // Out@GRAD@GRAD: ddy
+    op->SetOutput("DDOut", InputGrad(framework::GradVarName("Out")));
+    return std::unique_ptr<::paddle::framework::OpDesc>(op);
+  }
+};
+
 }  // namespace operators
 }  // namespace paddle
 
@@ -727,6 +745,7 @@ namespace plat = paddle::platform;
 FOR_EACH_ACTIVATION_OP(REGISTER_ACTIVATION_OP);
 FOR_EACH_ACTIVATION_OP(REGISTER_ACTIVATION_CPU_KERNEL);
 
+/* ==========================    relu register  ============================= */
 REGISTER_OPERATOR(
     relu, ops::ActivationOp, ops::ReluOpMaker, ops::ActivationOpInferVarType,
     ops::ActivationGradOpDescMaker<ops::ReluGradFunctor<float>::FwdDeps()>,
@@ -734,7 +753,9 @@ REGISTER_OPERATOR(
 REGISTER_OPERATOR(relu_grad, ops::ActivationOpGrad,
                   paddle::framework::SingleOpInplaceInToOut,
                   ops::ReluDoubleGradMaker);
-REGISTER_OPERATOR(relu_grad_grad, ops::ActivationOpDoubleGrad);
+REGISTER_OPERATOR(
+    relu_grad_grad,
+    ops::ActivationOpDoubleGrad<ops::ReluGradFunctor<float>::FwdDeps()>);
 
 REGISTER_ACTIVATION_CPU_KERNEL(relu, Relu, ReluFunctor, ReluGradFunctor);
 
@@ -746,7 +767,9 @@ REGISTER_OP_CPU_KERNEL(
                                     ops::ReluGradGradFunctor<double>>,
     ops::ActivationDoubleGradKernel<plat::CPUDeviceContext,
                                     ops::ReluGradGradFunctor<plat::float16>>);
+/* ========================================================================== */
 
+/* ======================== leaky relu register  ============================ */
 REGISTER_OPERATOR(
     leaky_relu, ops::ActivationOp, ops::LeakyReluOpMaker,
     ops::ActivationOpInferVarType,
@@ -755,7 +778,10 @@ REGISTER_OPERATOR(
 REGISTER_OPERATOR(leaky_relu_grad, ops::ActivationOpGrad,
                   paddle::framework::SingleOpInplaceInToOut,
                   ops::LeakyReluDoubleGradMaker);
-REGISTER_OPERATOR(leaky_relu_grad_grad, ops::LeakyReluDoubleGrad);
+REGISTER_OPERATOR(
+    leaky_relu_grad_grad,
+    ops::ActivationOpDoubleGrad<ops::LeakyReluGradFunctor<float>::FwdDeps()>);
+
 REGISTER_ACTIVATION_CPU_KERNEL(leaky_relu, LeakyRelu, LeakyReluFunctor,
                                LeakyReluGradFunctor);
 REGISTER_OP_CPU_KERNEL(
@@ -766,3 +792,30 @@ REGISTER_OP_CPU_KERNEL(
                                     ops::LeakyReluGradGradFunctor<double>>,
     ops::ActivationDoubleGradKernel<
         plat::CPUDeviceContext, ops::LeakyReluGradGradFunctor<plat::float16>>);
+/* ========================================================================== */
+
+/* ==========================   square register  ============================ */
+REGISTER_OPERATOR(
+    square, ops::ActivationOp, ops::SquareOpMaker,
+    ops::ActivationOpInferVarType,
+    ops::ActivationGradOpDescMaker<ops::SquareGradFunctor<float>::FwdDeps()>,
+    paddle::framework::SingleOpInplaceInToOut);
+REGISTER_OPERATOR(square_grad, ops::ActivationOpGrad,
+                  paddle::framework::SingleOpInplaceInToOut,
+                  ops::SquareDoubleGradMaker);
+REGISTER_OPERATOR(
+    square_grad_grad,
+    ops::ActivationOpDoubleGrad<ops::SquareGradGradFunctor<float>::FwdDeps()>);
+
+REGISTER_ACTIVATION_CPU_KERNEL(square, Square, SquareFunctor,
+                               SquareGradFunctor);
+
+REGISTER_OP_CPU_KERNEL(
+    square_grad_grad,
+    ops::SquareDoubleGradKernel<plat::CPUDeviceContext,
+                                ops::SquareGradGradFunctor<float>>,
+    ops::SquareDoubleGradKernel<plat::CPUDeviceContext,
+                                ops::SquareGradGradFunctor<double>>,
+    ops::SquareDoubleGradKernel<plat::CPUDeviceContext,
+                                ops::SquareGradGradFunctor<plat::float16>>);
+/* ========================================================================== */

paddle/fluid/operators/activation_op.cu

@@ -33,6 +33,7 @@ namespace plat = paddle::platform;
 
 FOR_EACH_ACTIVATION_OP(REGISTER_ACTIVATION_CUDA_KERNEL);
 
+/* ======================== leaky relu register  ============================ */
 REGISTER_ACTIVATION_CUDA_KERNEL(leaky_relu, LeakyRelu, LeakyReluFunctor,
                                 LeakyReluGradFunctor);
 
@@ -44,7 +45,9 @@ REGISTER_OP_CUDA_KERNEL(
                                     ops::LeakyReluGradGradFunctor<double>>,
     ops::ActivationDoubleGradKernel<
         plat::CUDADeviceContext, ops::LeakyReluGradGradFunctor<plat::float16>>);
+/* ========================================================================== */
 
+/* ===========================    relu register  ============================ */
 REGISTER_ACTIVATION_CUDA_KERNEL(relu, Relu, ReluFunctor, ReluGradFunctor);
 
 REGISTER_OP_CUDA_KERNEL(
@@ -55,3 +58,18 @@ REGISTER_OP_CUDA_KERNEL(
                                     ops::ReluGradGradFunctor<double>>,
     ops::ActivationDoubleGradKernel<plat::CUDADeviceContext,
                                     ops::ReluGradGradFunctor<plat::float16>>);
+/* ========================================================================== */
+
+/* ===========================  square register  ============================ */
+REGISTER_ACTIVATION_CUDA_KERNEL(square, Square, SquareFunctor,
+                                SquareGradFunctor);
+
+REGISTER_OP_CUDA_KERNEL(
+    square_grad_grad,
+    ops::SquareDoubleGradKernel<paddle::platform::CUDADeviceContext,
+                                ops::SquareGradGradFunctor<float>>,
+    ops::SquareDoubleGradKernel<paddle::platform::CUDADeviceContext,
+                                ops::SquareGradGradFunctor<double>>,
+    ops::SquareDoubleGradKernel<plat::CUDADeviceContext,
+                                ops::SquareGradGradFunctor<plat::float16>>);
+/* ========================================================================== */

paddle/fluid/operators/activation_op.h

 /* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
+
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
 You may obtain a copy of the License at
@@ -1358,6 +1359,90 @@ struct LeakyReluGradGradFunctor : public BaseActivationFunctor<T> {
   static constexpr ActBwdOpFwdDeps FwdDeps() { return kDepX; }
 };
 
+template <typename T>
+struct SquareGradGradFunctor : public BaseActivationFunctor<T> {
+  template <typename Device>
+  void operator()(const Device& dev, const framework::Tensor* X,
+                  const framework::Tensor* ddX, framework::Tensor* ddOut,
+                  const framework::Tensor* dOut, framework::Tensor* dX) const {
+    auto* d = dev.eigen_device();
+    auto ddx = framework::EigenVector<T>::Flatten(detail::Ref(ddX));
+    auto x = framework::EigenVector<T>::Flatten(detail::Ref(X));
+    if (ddOut) {
+      auto ddout = framework::EigenVector<T>::Flatten(detail::Ref(ddOut));
+      ddout.device(*d) = ddx * static_cast<T>(2) * x;
+    }
+    if (dX) {
+      auto dx = framework::EigenVector<T>::Flatten(detail::Ref(dX));
+      auto dout = framework::EigenVector<T>::Flatten(detail::Ref(dOut));
+      dx.device(*d) = ddx * static_cast<T>(2) * dout;
+    }
+  }
+  static constexpr ActBwdOpFwdDeps FwdDeps() { return kDepX; }
+};
+
+// TODO(dengkaipeng): double gradient calculation for Square/Sqrt need
+// DOut(dy) as input(not output), tensor extraction is different from
+// others. Impliment extraction kernel seperately here.
+inline void ExtractDoubleGradTensorWithInputDOut(
+    const framework::ExecutionContext& ctx, const framework::Tensor** X,
+    const framework::Tensor** ddX, framework::Tensor** dX,
+    const framework::Tensor** dOut, framework::Tensor** ddOut) {
+  // extract ddX(output), ddOut(input)
+  auto ddx_var = ctx.InputVar("DDX");
+  auto ddo_var = ctx.OutputVar("DDOut");
+  PADDLE_ENFORCE(ddx_var != nullptr,
+                 "Cannot get input Variable Out, variable name = %s",
+                 ctx.op().Input("DDX"));
+  *ddX = ctx.Input<framework::Tensor>("DDX");
+  if (ddo_var) {
+    *ddOut = ctx.Output<framework::Tensor>("DDOut");
+  }
+  PADDLE_ENFORCE(*ddX != nullptr,
+                 "Cannot get output tensor DDX, variable name = %s",
+                 ctx.op().Output("DDX"));
+
+  // extract x(input), dx(output)
+  auto x_var = ctx.InputVar("X");
+  PADDLE_ENFORCE(x_var != nullptr,
+                 "Cannot get input Variable Out, variable name = %s",
+                 ctx.op().Input("X"));
+  auto dx_var = ctx.OutputVar("DX");
+  *X = ctx.Input<framework::Tensor>("X");
+  if (dx_var) {
+    *dX = ctx.Output<framework::Tensor>("DX");
+  }
+
+  // extract dOut(input)
+  auto dout_var = ctx.InputVar("DOut");
+  if (dout_var) {
+    *dOut = ctx.Input<framework::Tensor>("DOut");
+  }
+}
+
+template <typename DeviceContext, typename Functor>
+class SquareDoubleGradKernel
+    : public framework::OpKernel<typename Functor::ELEMENT_TYPE> {
+ public:
+  using T = typename Functor::ELEMENT_TYPE;
+  void Compute(const framework::ExecutionContext& ctx) const override {
+    const framework::Tensor *X, *ddX, *dOut;
+    X = ddX = dOut = nullptr;
+    framework::Tensor *dX, *ddOut;
+    dX = ddOut = nullptr;
+
+    ExtractDoubleGradTensorWithInputDOut(ctx, &X, &ddX, &dX, &dOut, &ddOut);
+
+    dX->mutable_data<T>(X->dims(), ctx.GetPlace());
+    ddOut->mutable_data<T>(ctx.GetPlace());
+
+    auto& place = ctx.template device_context<DeviceContext>();
+
+    Functor functor;
+    functor(place, X, ddX, ddOut, dOut, dX);
+  }
+};
+
 }  // namespace operators
 }  // namespace paddle
 
@@ -1381,7 +1466,6 @@ struct LeakyReluGradGradFunctor : public BaseActivationFunctor<T> {
   __macro(round, Round, RoundFunctor, ZeroGradFunctor);                       \
   __macro(reciprocal, Reciprocal, ReciprocalFunctor, ReciprocalGradFunctor);  \
   __macro(log, Log, LogFunctor, LogGradFunctor);                              \
-  __macro(square, Square, SquareFunctor, SquareGradFunctor);                  \
   __macro(brelu, BRelu, BReluFunctor, BReluGradFunctor);                      \
   __macro(soft_relu, SoftRelu, SoftReluFunctor, SoftReluGradFunctor);         \
   __macro(pow, Pow, PowFunctor, PowGradFunctor);                              \

python/paddle/fluid/tests/unittests/test_nn_grad.py

@@ -115,5 +115,29 @@ class TestConvDoubleGradCheck(unittest.TestCase):
                 self.func(p)
 
 
+class TestSquareDoubleGradCheck(unittest.TestCase):
+    @prog_scope()
+    def func(self, place):
+        # the shape of input variable shoule be clearly specified, not inlcude -1.
+        shape = [17, 23]
+        eps = 0.005
+        dtype = np.float64
+
+        x = layers.data('x', shape, False, dtype)
+        x.persistable = True
+        y = layers.square(x)
+        x_arr = np.random.uniform(-1, 1, shape).astype(dtype)
+
+        gradient_checker.double_grad_check(
+            [x], y, x_init=x_arr, place=place, eps=eps)
+
+    def test_grad(self):
+        places = [fluid.CPUPlace()]
+        if core.is_compiled_with_cuda():
+            places.append(fluid.CUDAPlace(0))
+        for p in places:
+            self.func(p)
+
+
 if __name__ == "__main__":
     unittest.main()