You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Paddle/paddle/fluid/operators/tree_conv_op.cc

227 lines
9.3 KiB

// 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
//
// 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/fluid/operators/tree_conv_op.h"
#include <memory>
#include <string>
namespace paddle {
namespace operators {
class TreeConvOpMaker : public framework::OpProtoAndCheckerMaker {
public:
void Make() override {
AddInput("NodesVector",
"(Tensor) The feature vector of every node on the tree. "
"The shape of the feature vector must be "
"[max_tree_node_size, feature_size].");
AddInput("EdgeSet",
"(Tensor) The Edges of Tree. The edge must be directional. "
"The shape of the edge set must be [max_tree_node_size, 2].");
AddInput("Filter",
"(Tensor) The feature detector. "
"The shape of the filter is "
"[feature_size, 3, output_size, num_filters].");
AddOutput("Out",
"(Tensor) The feature vector of subtrees. "
"The shape of the output tensor is [max_tree_node_size, "
"output_size, num_filters]. "
"The output tensor could be a new feature "
"vector for next tree convolution layers.");
AddAttr<int>("max_depth",
"(int, default: 2) The depth of feature detector.")
.SetDefault(2)
.GreaterThan(1);
AddComment(R"DOC(
**Tree-Based Convolution Operator**
Tree-Based Convolution is a kind of convolution based on tree structure.
Tree-Based Convolution is a part of Tree-Based Convolution Neural Network(TBCNN),
which is used to classify tree structures, such as Abstract Syntax Tree.
Tree-Based Convolution proposed a kind of data structure called continuous binary tree,
which regards multiway tree as binary tree.
The paper of Tree-Based Convolution Operator is here:
https://arxiv.org/abs/1409.5718v1
)DOC");
}
};
class TreeConvOp : public framework::OperatorWithKernel {
public:
using framework::OperatorWithKernel::OperatorWithKernel;
void InferShape(framework::InferShapeContext *ctx) const override {
OP_INOUT_CHECK(ctx->HasInput("NodesVector"), "Input", "NodesVector",
"TreeConv");
OP_INOUT_CHECK(ctx->HasInput("Filter"), "Input", "Filter", "TreeConv");
OP_INOUT_CHECK(ctx->HasInput("EdgeSet"), "Input", "EdgeSet", "TreeConv");
OP_INOUT_CHECK(ctx->HasOutput("Out"), "Output", "Out", "TreeConv");
auto edge_dims = ctx->GetInputDim("EdgeSet");
auto vector_dims = ctx->GetInputDim("NodesVector");
auto filter_dims = ctx->GetInputDim("Filter");
if (ctx->IsRuntime()) {
PADDLE_ENFORCE_EQ(edge_dims[2], 2,
platform::errors::InvalidArgument(
"Input(EdgeSet) dim[2] should be 2. "
"But received Input(EdgeSet) dim[2] is %d.",
edge_dims[2]));
} else {
if (edge_dims[2] != -1) {
PADDLE_ENFORCE_EQ(edge_dims[2], 2,
platform::errors::InvalidArgument(
"Input(EdgeSet) dim[2] should be 2. "
"But received Input(EdgeSet) dim[2] is %d.",
edge_dims[2]));
}
}
PADDLE_ENFORCE_EQ(edge_dims.size(), 3,
platform::errors::InvalidArgument(
"The dimension of EdgeSet Tensor should be 3. "
"But received the dimension of EdgeSet Tensor is %d.",
edge_dims.size()));
PADDLE_ENFORCE_EQ(
vector_dims.size(), 3,
platform::errors::InvalidArgument(
"The dimension of NodesVector Tensor should be 3. "
"But received the dimension of NodesVector Tensor is %d.",
vector_dims.size()));
PADDLE_ENFORCE_EQ(filter_dims.size(), 4,
platform::errors::InvalidArgument(
"The dimension of Filter Tensor should be 4. "
"But received the dimension of Filter Tensor is %d.",
filter_dims.size()));
if (ctx->IsRuntime()) {
PADDLE_ENFORCE_EQ(filter_dims[1], 3,
platform::errors::InvalidArgument(
"Input(Filter) dim[1] should be 3. "
"But received Input(Filter) dim[1] is %d.",
filter_dims[1]));
PADDLE_ENFORCE_EQ(
filter_dims[0], vector_dims[2],
platform::errors::InvalidArgument(
"Input(Filter) dim[0] must equal to Input(NodesVector) dim[2]. "
"But received Input(Filter) dim[0] = %d, Input(NodesVector) "
"dim[2] = %d.",
filter_dims[0], vector_dims[2]));
} else {
if (filter_dims[1] != -1) {
PADDLE_ENFORCE_EQ(filter_dims[1], 3,
platform::errors::InvalidArgument(
"Input(Filter) dim[1] should be 3. "
"But received Input(Filter) dim[1] is %d.",
filter_dims[1]));
}
if (filter_dims[0] != -1 && vector_dims[2] != -1) {
PADDLE_ENFORCE_EQ(
filter_dims[0], vector_dims[2],
platform::errors::InvalidArgument(
"Input(Filter) dim[0] must equal to Input(NodesVector) dim[2]. "
"But received Input(Filter) dim[0] = %d, Input(NodesVector) "
"dim[2] = %d.",
filter_dims[0], vector_dims[2]));
}
}
auto output_dims = framework::make_ddim(
{vector_dims[0], vector_dims[1], filter_dims[2], filter_dims[3]});
ctx->SetOutputDim("Out", output_dims);
}
protected:
framework::OpKernelType GetExpectedKernelType(
const framework::ExecutionContext &ctx) const override {
return framework::OpKernelType(
OperatorWithKernel::IndicateVarDataType(ctx, "NodesVector"),
ctx.device_context());
}
};
template <typename T>
class TreeConvGradOpMaker : public framework::SingleGradOpMaker<T> {
public:
using framework::SingleGradOpMaker<T>::SingleGradOpMaker;
protected:
void Apply(GradOpPtr<T> op) const override {
op->SetType("tree_conv_grad");
op->SetInput(framework::GradVarName("Out"), this->OutputGrad("Out"));
op->SetInput("Filter", this->Input("Filter"));
op->SetInput("EdgeSet", this->Input("EdgeSet"));
op->SetInput("NodesVector", this->Input("NodesVector"));
op->SetOutput(framework::GradVarName("NodesVector"),
this->InputGrad("NodesVector"));
op->SetOutput(framework::GradVarName("Filter"), this->InputGrad("Filter"));
op->SetAttrMap(this->Attrs());
}
};
class TreeConvGradOp : public framework::OperatorWithKernel {
public:
using framework::OperatorWithKernel::OperatorWithKernel;
void InferShape(framework::InferShapeContext *ctx) const override {
OP_INOUT_CHECK(ctx->HasInput("Filter"), "Input", "Filter", "grad_TreeConv");
OP_INOUT_CHECK(ctx->HasInput("EdgeSet"), "Input", "EdgeSet",
"grad_TreeConv");
OP_INOUT_CHECK(ctx->HasInput("NodesVector"), "Input", "NodesVector",
"grad_TreeConv");
OP_INOUT_CHECK(ctx->HasInput(framework::GradVarName("Out")), "Input",
framework::GradVarName("Out"), "grad_TreeConv");
OP_INOUT_CHECK(ctx->HasOutput(framework::GradVarName("NodesVector")),
"Output", framework::GradVarName("NodesVector"),
"grad_TreeConv");
OP_INOUT_CHECK(ctx->HasOutput(framework::GradVarName("Filter")), "Output",
framework::GradVarName("Filter"), "grad_TreeConv");
auto vectors_dims = ctx->GetInputDim("NodesVector");
auto filter_dims = ctx->GetInputDim("Filter");
if (ctx->HasOutput(framework::GradVarName("Filter"))) {
ctx->SetOutputDim(framework::GradVarName("Filter"), filter_dims);
}
if (ctx->HasOutput(framework::GradVarName("NodesVector"))) {
ctx->SetOutputDim(framework::GradVarName("NodesVector"), vectors_dims);
}
}
protected:
framework::OpKernelType GetExpectedKernelType(
const framework::ExecutionContext &ctx) const override {
return framework::OpKernelType(
OperatorWithKernel::IndicateVarDataType(ctx, "NodesVector"),
ctx.device_context());
}
};
} // namespace operators
} // namespace paddle
namespace ops = paddle::operators;
REGISTER_OPERATOR(tree_conv, ops::TreeConvOp, ops::TreeConvOpMaker,
ops::TreeConvGradOpMaker<paddle::framework::OpDesc>,
ops::TreeConvGradOpMaker<paddle::imperative::OpBase>);
REGISTER_OPERATOR(tree_conv_grad, ops::TreeConvGradOp);
REGISTER_OP_CPU_KERNEL(
tree_conv, ops::TreeConvKernel<paddle::platform::CPUDeviceContext, float>,
ops::TreeConvKernel<paddle::platform::CPUDeviceContext, double>);
REGISTER_OP_CPU_KERNEL(
tree_conv_grad,
ops::TreeConvGradKernel<paddle::platform::CPUDeviceContext, float>,
ops::TreeConvGradKernel<paddle::platform::CPUDeviceContext, double>);