(this PR is just for my personal self-learning) [PaddlePaddle Hackathon 2 No.22] add paddle.index_add to Paddle

paddle/fluid/operators/index_add_op.cc

@@ -0,0 +1,115 @@
+/*Copyright (c) 2022 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
+    http://www.apache.org/licenses/LICENSE-2.0
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include <memory>
+#include "paddle/fluid/framework/infershape_utils.h"
+#include "paddle/fluid/framework/op_registry.h"
+#include "paddle/fluid/framework/op_version_registry.h"
+#include "paddle/phi/core/infermeta_utils.h"
+#include "paddle/phi/infermeta/binary.h"
+
+namespace paddle {
+namespace operators {
+
+class IndexAddOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext& ctx) const override {
+    return framework::OpKernelType(
+        OperatorWithKernel::IndicateVarDataType(ctx, "X"), ctx.GetPlace());
+  }
+};
+
+class IndexAddOpMaker : public framework::OpProtoAndCheckerMaker {
+ public:
+  void Make() override {
+    AddInput("X",
+             "(Tensor, default input Tensor<float>), "
+             "the input feature data of IndexAddOp, dtype should be"
+             "int32, int64, float16, float32, float64.");
+    AddInput("Index",
+             "(Tensor, default 1-d Tensor<int>), "
+             "the 1-D tensor containing the indices to index, "
+             "dtype should be int32, int64");
+    AddAttr<int>("axis",
+                 "(int, default 0), "
+                 "the dimension in which we index.")
+        .SetDefault(0);
+    AddAttr<float>("added_value",
+                   "(float, default 0.0f) The value to add.")
+        .SetDefault(0.0f);
+    AddOutput("Out",
+              "(Tensor, default Tensor<float>),"
+              " the output of  IndexAddOp, whose dtype and shape is the same as X.");
+    AddComment(R"DOC(
+index_add operator.
+Add the elements of the input tensor with value
+by selecting the indices in the order given in 'index'
+on the axis 'axis'.
+
+This operator also supports inplace modification.
+        )DOC");
+  }
+};
+
+template <typename T>
+class IndexAddGradOpMaker : public framework::SingleGradOpMaker<T> {
+ public:
+  using framework::SingleGradOpMaker<T>::SingleGradOpMaker;
+
+  void Apply(GradOpPtr<T> op) const override {
+    op->SetType("index_add_grad");
+    op->SetInput(framework::GradVarName("Out"), this->OutputGrad("Out"));
+    op->SetOutput(framework::GradVarName("X"), this->InputGrad("X"));
+    op->SetAttrMap(this->Attrs());
+  }
+};
+
+class IndexAddGradOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+ protected:
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext& ctx) const override {
+    return framework::OpKernelType(
+      OperatorWithKernel::IndicateVarDataType(
+          ctx, framework::GradVarName("Out")), ctx.GetPlace());
+  }
+};
+
+DECLARE_INPLACE_OP_INFERER(IndexAddInplaceInferer, {"X", "Out"});
+DECLARE_INPLACE_OP_INFERER(IndexAddGradInplaceInferer,
+                           {framework::GradVarName("Out"),
+                            framework::GradVarName("X")});
+// DECLARE_NO_NEED_BUFFER_VARS_INFERER(IndexAddGradNoNeedBufferVarsInferer, "X");
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+DECLARE_INFER_SHAPE_FUNCTOR(index_add, IndexAddInferShapeFunctor,
+                            PD_INFER_META(phi::IndexAddInferMeta));
+
+REGISTER_OPERATOR(index_add, ops::IndexAddOp, ops::IndexAddOpMaker,
+                  ops::IndexAddGradOpMaker<paddle::framework::OpDesc>,
+                  ops::IndexAddGradOpMaker<paddle::imperative::OpBase>,
+                  ops::IndexAddInplaceInferer, IndexAddInferShapeFunctor);
+
+DECLARE_INFER_SHAPE_FUNCTOR(index_add_grad, IndexAddGradInferShapeFunctor,
+                            PD_INFER_META(phi::IndexAddGradInferMeta));
+
+REGISTER_OPERATOR(index_add_grad, ops::IndexAddGradOp,
+                  ops::IndexAddGradInplaceInferer,
+                  // ops::IndexAddGradNoNeedBufferVarsInferer,
+                  IndexAddGradInferShapeFunctor);

paddle/phi/infermeta/binary.cc

@@ -1181,6 +1181,51 @@ void HuberLossInferMeta(const MetaTensor& input,
   out->share_lod(input);
 }
 
+void IndexAddInferMeta(const MetaTensor& x,
+                       const MetaTensor& index,
+                       int axis,
+                       float added_value,
+                       MetaTensor* output) {
+  auto input_dim = x.dims();
+  auto index_dim = index.dims();
+
+  PADDLE_ENFORCE_EQ(
+      axis < input_dim.size() && axis >= - input_dim.size(),
+      true,
+      phi::errors::OutOfRange(
+          "Attr(axis) is out of range, It's expected "
+          "to be in range of [-%d, %d). But received Attr(axis) = %d.",
+          input_dim.size(),
+          input_dim.size(),
+          axis));
+
+  PADDLE_ENFORCE_EQ(
+      // index_dim.size() == 1 || (index_dim.size() == 2 && index_dim[1] == 1),
+      index_dim.size() == 1 && index_dim[0] > 0,
+      true,
+      phi::errors::InvalidArgument(
+          "The 'shape' of Input(Index) must be 1-D tensor. "
+          "But received: the 'shape' of Input(Index) is [%s], "
+          "the dimension of Input(Index) is [%d].",
+          index_dim,
+          index_dim.size()));
+
+  output->set_dims(x.dims());
+  output->set_dtype(x.dtype());
+  output->set_layout(x.layout());
+  output->share_lod(x);
+}
+
+void IndexAddGradInferMeta(const MetaTensor& out_grad,
+                           int axis,
+                           float added_value,
+                           MetaTensor* x_grad) {
+  x_grad->set_dims(out_grad.dims());
+  x_grad->set_dtype(out_grad.dtype());
+  x_grad->set_layout(out_grad.layout());
+  x_grad->share_lod(out_grad);
+}
+
 void IndexSampleInferMeta(const MetaTensor& x,
                           const MetaTensor& y,
                           MetaTensor* out,

paddle/phi/infermeta/binary.h

@@ -184,6 +184,17 @@ void HuberLossInferMeta(const MetaTensor& input_meta,
                         MetaTensor* residual,
                         MetaConfig config = MetaConfig());
 
+void IndexAddInferMeta(const MetaTensor& x,
+                       const MetaTensor& index,
+                       int axis,
+                       float added_value,
+                       MetaTensor* output);
+
+void IndexAddGradInferMeta(const MetaTensor& out_grad,
+                           int axis,
+                           float added_value,
+                           MetaTensor* x_grad);
+
 void IndexSampleInferMeta(const MetaTensor& x,
                           const MetaTensor& y,
                           MetaTensor* out,

paddle/phi/kernels/cpu/index_add_grad_kernel.cc

@@ -0,0 +1,41 @@
+// Copyright (c) 2022 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
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "paddle/phi/kernels/index_add_grad_kernel.h"
+#include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/core/utils/data_type.h"
+#include "paddle/phi/kernels/copy_kernel.h"
+
+namespace phi {
+
+template <typename T, typename Context>
+void IndexAddGradKernel(const Context& dev_ctx,
+                        const DenseTensor& out_grad,
+                        int axis,
+                        float added_value,
+                        DenseTensor* x_grad) {
+  phi::Copy(dev_ctx, out_grad, dev_ctx.GetPlace(), false, x_grad);
+}
+
+}  // namespace phi
+
+PD_REGISTER_KERNEL(index_add_grad,
+                   CPU,
+                   ALL_LAYOUT,
+                   phi::IndexAddGradKernel,
+                   float,
+                   phi::dtype::float16,
+                   double,
+                   int,
+                   int64_t) {}

paddle/phi/kernels/cpu/index_add_impl.h

@@ -0,0 +1,86 @@
+// Copyright (c) 2022 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
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#pragma once
+
+#include "paddle/phi/core/dense_tensor.h"
+#include "paddle/phi/kernels/copy_kernel.h"
+#include "paddle/phi/kernels/funcs/blas/blas.h"
+#include "paddle/phi/kernels/funcs/eigen/common.h"
+#include "paddle/phi/kernels/funcs/math_function.h"
+
+namespace phi {
+template <typename Context, typename T, typename IndexT = int>
+void IndexAddInner(const Context& ctx,
+                   const DenseTensor& index,
+                   DenseTensor* output,
+                   int axis,
+                   T added_val) {
+  auto output_dim = output->dims();
+  auto output_dim_size = output_dim.size();
+  auto index_size = index.dims()[0];
+
+  DenseTensor index_cpu_copy;
+  if (!paddle::platform::is_cpu_place(index.place())) {
+    phi::Copy(ctx, index, phi::CPUPlace(), true, &index_cpu_copy);
+  }
+  const IndexT* index_data = paddle::platform::is_cpu_place(index.place())
+                                 ? index.data<IndexT>()
+                                 : index_cpu_copy.data<IndexT>();
+
+  auto slice_size = 1;
+  for (auto i = axis + 1; i < output_dim_size; i++) {
+    slice_size *= output_dim[i];
+  }
+
+  auto outer_nums = 1;
+  for (auto i = 0; i < axis; i++) {
+    outer_nums *= output_dim[i];
+  }
+
+  for (int i = 0; i < index_size; i++) {
+    PADDLE_ENFORCE_GE(
+        index_data[i],
+        0,
+        phi::errors::InvalidArgument(
+            "Variable value (index) of OP(index_add) "
+            "expected >= 0 and < %ld, but got %ld. Please check input "
+            "value.",
+            output_dim[axis],
+            index_data[i]));
+    PADDLE_ENFORCE_LT(
+        index_data[i],
+        output_dim[axis],
+        phi::errors::InvalidArgument(
+            "Variable value (index) of OP(index_add) "
+            "expected >= 0 and < %ld, but got %ld. Please check input "
+            "value.",
+            output_dim[axis],
+            index_data[i]));
+  }
+
+  output->Resize(phi::make_ddim({outer_nums, output_dim[axis], slice_size}));
+
+  auto output_tensor = EigenTensor<T, 3>::From(*output);
+  auto& place = *ctx.eigen_device();
+  for (auto j = 0; j < index_size; j++) {
+    IndexT index_value = index_data[j];
+    auto output_t = output_tensor.chip(index_value, 1);
+    // output_t.device(place) = output_t.constant(fill_val);
+    output_t.device(place) += output_t.constant(added_val);
+  }
+  output->Resize(output_dim);
+}
+
+}  // namespace phi

paddle/phi/kernels/cpu/index_add_kernel.cc

@@ -0,0 +1,66 @@
+// Copyright (c) 2022 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
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "paddle/phi/kernels/index_add_kernel.h"
+#include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/core/utils/data_type.h"
+#include "paddle/phi/kernels/copy_kernel.h"
+#include "paddle/phi/kernels/cpu/index_add_impl.h"
+#include "paddle/phi/kernels/funcs/eigen/common.h"
+
+namespace phi {
+
+template <typename T, typename Context>
+void IndexAddKernel(const Context& dev_ctx,
+                    const DenseTensor& x,
+                    const DenseTensor& index,
+                    int axis,
+                    float added_value,
+                    DenseTensor* output) {
+  phi::Copy(dev_ctx, x, dev_ctx.GetPlace(), false, output);
+  if (axis < 0) {
+    axis += x.dims().size();
+  }
+  const auto& index_type = index.dtype();
+
+  bool index_type_match =
+      index_type == phi::DataType::INT32 || index_type == phi::DataType::INT64;
+  PADDLE_ENFORCE_EQ(index_type_match,
+                    true,
+                    phi::errors::InvalidArgument(
+                        "Input(Index) holds the wrong type, it holds %s, but "
+                        "desires to be %s or %s",
+                        index_type,
+                        phi::DataType::INT32,
+                        phi::DataType::INT64));
+
+  auto added_val = static_cast<T>(added_value);
+  if (index_type == phi::DataType::INT32) {
+    IndexAddInner<Context, T, int>(dev_ctx, index, output, axis, added_val);
+  } else if (index_type == phi::DataType::INT64) {
+    IndexAddInner<Context, T, int64_t>(dev_ctx, index, output, axis, added_val);
+  }
+}
+
+}  // namespace phi
+
+PD_REGISTER_KERNEL(index_add,
+                   CPU,
+                   ALL_LAYOUT,
+                   phi::IndexAddKernel,
+                   float,
+                   phi::dtype::float16,
+                   double,
+                   int,
+                   int64_t) {}