Shape \
  ExpandDims \
  OnesLike

${generate} \
  --category=math \
  Mul \
  Conj \
  AddV2 \
  MatMul \
  Neg \
  Sum \
  Sub \
  Div \
  DivNoNan \
  Exp \
  Sqrt \
  SqrtGrad \
  Log1p

${generate} \
  --category=nn \
  SparseSoftmaxCrossEntropyWithLogits \
  ReluGrad \
  Relu \
  BiasAdd \
  BiasAddGrad

${generate} \
  --category=resource_variable \

  int num_retvals = 1;
  return op_ptr->Execute(absl::MakeSpan(y, 1), &num_retvals);
}

// Op: SqrtGrad()
// Summary: Computes the gradient for the sqrt of `x` wrt its input.
//
// Description:
//   Specifically, `grad = dy * 0.5 / y`, where `y = sqrt(x)`, and `dy`
//   is the corresponding input gradient.
Status SqrtGrad(AbstractContext* ctx, AbstractTensorHandle* const y,

            AbstractTensorHandle** y, const char* name = nullptr,
            const char* raw_device_name = nullptr);

// Computes the gradient for the sqrt of `x` wrt its input.
Status SqrtGrad(AbstractContext* ctx, AbstractTensorHandle* const y,
                AbstractTensorHandle* const dy, AbstractTensorHandle** z,
                const char* name = nullptr,
                const char* raw_device_name = nullptr);

using tensorflow::ops::Div;
using tensorflow::ops::DivNoNan;
using tensorflow::ops::MatMul;
using tensorflow::ops::Mul;
using tensorflow::ops::Neg;
using tensorflow::ops::OnesLike;
using tensorflow::ops::SqrtGrad;

namespace tensorflow {
namespace gradients {
namespace {

static Status SafeConj(AbstractContext* ctx, AbstractTensorHandle* const input,
                       AbstractTensorHandle** output, const char* name) {

  )>;

//===----------------------------------------------------------------------===//
// Sqrt op patterns.
//===----------------------------------------------------------------------===//

// SqrtGrad(y, dy) = dy * 0.5 / y
def LowerSqrtGradOp : Pat<
  (TF_SqrtGradOp $y, $dy),
  (TF_DivOp
    (TF_MulOp $dy, (TF_ConstOp (GetScalarOfFloatType<"0.5"> $dy))),
    $y
  )>;

REGISTER_GRADIENT_OP("Square", SquareGrad);

Status SqrtGrad(const Scope& scope, const Operation& op,
                const std::vector<Output>& grad_inputs,
                std::vector<Output>* grad_outputs) {
  // Use the built-in operator.
  grad_outputs->push_back(
      internal::SqrtGrad(scope, op.output(0), grad_inputs[0]));
  return scope.status();
}
REGISTER_GRADIENT_OP("Sqrt", SqrtGrad);

            "BitwiseAnd", "BitwiseOr", "BitwiseXor", "LeftShift", "RightShift",
            "LogicalAnd", "LogicalOr", "Mod", "Maximum", "Minimum", "RealDiv",
            "ReciprocalGrad", "RsqrtGrad", "SqrtGrad", "TruncateDiv",
            "TruncateMod", "Equal", "NotEqual", "Greater", "GreaterEqual",
            "Less", "LessEqual", "SigmoidGrad", "SoftplusGrad", "SoftsignGrad",

  // CHECK: %[[RET:.*]] = "tf.Div"(%[[MUL]], %arg0) : (tensor<*xcomplex<f32>>, tensor<*xcomplex<f32>>) -> tensor<*xcomplex<f32>>

  %0 = "tf.SqrtGrad"(%arg0, %arg1) : (tensor<*xcomplex<f32>>, tensor<*xcomplex<f32>>) -> tensor<*xcomplex<f32>>
  // CHECK: return %[[RET]]
  func.return %0 : tensor<*xcomplex<f32>>
}

    static bool isCompatibleReturnTypes(TypeRange inferred, TypeRange actual) {
      return ArraysAreCastCompatible(inferred, actual);
    }
  }];
}

def TF_SqrtGradOp : TF_Op<"SqrtGrad", [Pure, TF_SameOperandsAndResultTypeResolveRef]> {
  let summary = "Computes the gradient for the sqrt of `x` wrt its input.";

  let description = [{