sequence_pool_op.cc 4.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.

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. */

15
#include "paddle/operators/sequence_pool_op.h"
16 17 18 19

namespace paddle {
namespace operators {

20
class SequencePoolOp : public framework::OperatorWithKernel {
21 22 23 24
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

 protected:
Q
Qiao Longfei 已提交
25 26
  void InferShape(framework::InferShapeContextBase* ctx) const override {
    PADDLE_ENFORCE(ctx->HasInput("X"),
27
                   "Input(X) of SequencePoolOp should not be null.");
Q
Qiao Longfei 已提交
28
    PADDLE_ENFORCE(ctx->HasOutput("Out"),
29
                   "Output(Out) of SequencePoolOp should not be null.");
Q
Qiao Longfei 已提交
30
    ctx->SetOutputDim("Out", ctx->GetInputDim("X"));
31 32 33
  }
};

34
class SequencePoolOpMaker : public framework::OpProtoAndCheckerMaker {
35
 public:
36 37
  SequencePoolOpMaker(framework::OpProto* proto,
                      framework::OpAttrChecker* op_checker)
38
      : OpProtoAndCheckerMaker(proto, op_checker) {
L
Luo Tao 已提交
39 40 41 42 43
    AddInput("X",
             "A float LoDTensor, the variable-length input of SequencePoolOp");
    AddOutput(
        "Out",
        "A float LoDTensor, the variable-length output of SequencePoolOp.");
44 45 46 47 48
    AddAttr<int>(
        "strategy",
        "(int, default AVERAGE) the pooling strategy of SequencePoolOp.")
        .SetDefault(AVERAGE)
        .InEnum({AVERAGE, SUM, SQRT, MAX, LAST, FIRST});
49
    AddComment(R"DOC(
50 51
    SequencePoolOp pools features of all time-steps of each instance.

L
Luo Tao 已提交
52
    For a mini-batch of 3 variable lengths sentences, containing 2, 3, and 2 time-steps:
Q
Qiao Longfei 已提交
53

L
Luo Tao 已提交
54
    Assume X is a [7,M,N] float LoDTensor, and X->lod()[0] = [0, 2, 5, 7].
Q
Qiao Longfei 已提交
55
    Besides, for the sake of simplicity, we assume M=1 and N=1,
L
Luo Tao 已提交
56 57 58
    and the value of X = [[1, 3], [2, 4, 6], [5, 1]].

    Thus, Out is a [3,1,1] float LoDTensor, but Out->lod() is nullptr.
Q
Qiao Longfei 已提交
59
    And for different strategy, the value of Out is as follows:
L
Luo Tao 已提交
60 61 62

    - AVERAGE: [2, 4, 3], where 2=(1+3)/2, 4=(2+4+6)/3, 3=(5+1)/2
    - SUM: [4, 12, 6], where 4=1+3, 12=2+4+6, 6=5+1
Q
Qiao Longfei 已提交
63
    - SQRT: [2.82, 6.93, 4.24], where 2.82=(1+3)/sqrt(2),
L
Luo Tao 已提交
64 65 66 67
           6.93=(2+4+6)/sqrt(3), 4.24=(5+1)/sqrt(2)
    - MAX: [3, 6, 5], where 3=max(1,3), 6=max(2,4,6), 5=max(5,1)
    - LAST: [3, 6, 1], where 3=last(1,3), 6=last(2,4,6), 1=last(5,1)
    - FIRST: [1, 2, 5], where 1=first(1,3), 2=first(2,4,6), 5=first(5,1)
68 69 70 71
    )DOC");
  }
};

72
class SequencePoolGradOp : public framework::OperatorWithKernel {
73 74 75 76
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

 protected:
Q
Qiao Longfei 已提交
77 78 79 80 81 82
  void InferShape(framework::InferShapeContextBase* ctx) const override {
    PADDLE_ENFORCE(ctx->HasInput(framework::GradVarName("Out")),
                   "Gradient of Out should not be null.");
    PADDLE_ENFORCE(ctx->HasInput("X"), "The input X should not be null.");
    auto og_dims = ctx->GetInputDim(framework::GradVarName("Out"));
    auto x_dims = ctx->GetInputDim("X");
83 84
    PADDLE_ENFORCE_EQ(og_dims.size(), x_dims.size(),
                      "The rank of output grad must equal to Input(X).");
85
    for (int64_t i = 1; i < og_dims.size(); ++i) {
86 87
      PADDLE_ENFORCE_EQ(og_dims[i], x_dims[i], "The dimension mismatch.");
    }
Q
Qiao Longfei 已提交
88
    ctx->SetOutputDim(framework::GradVarName("X"), x_dims);
89 90 91 92 93 94 95
  }
};

}  // namespace operators
}  // namespace paddle

namespace ops = paddle::operators;
96 97
REGISTER_OP(sequence_pool, ops::SequencePoolOp, ops::SequencePoolOpMaker,
            sequence_pool_grad, ops::SequencePoolGradOp);
98
REGISTER_OP_CPU_KERNEL(
99
    sequence_pool, ops::SequencePoolKernel<paddle::platform::CPUPlace, float>);
100
REGISTER_OP_CPU_KERNEL(
101 102
    sequence_pool_grad,
    ops::SequencePoolGradKernel<paddle::platform::CPUPlace, float>);