cross_entropy_op.cu 5.5 KB
Newer Older
L
liaogang 已提交
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/framework/op_registry.h"
16
#include "paddle/operators/cross_entropy_op.h"
17
#include "paddle/platform/assert.h"
18
#include "paddle/platform/hostdevice.h"
19 20 21 22 23 24 25 26 27 28 29 30

namespace paddle {
namespace operators {

template <typename T>
__global__ void CrossEntropyKernel(T* Y, const T* X, const int* label,
                                   const int N, const int D) {
  // TOOD(qingqing) define CUDA_1D_KERNEL_LOOP macro in a common file.
  // CUDA_1D_KERNEL_LOOP(i, N) {
  for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < N;
       i += blockDim.x * gridDim.x) {
    PADDLE_ASSERT(label[i] >= 0 && label[i] < D);
31 32 33 34 35 36 37 38 39 40 41
    Y[i] = -tolerable_value(log(X[i * D + label[i]]));
  }
}

template <typename T>
__global__ void SoftCrossEntropyKernel(T* Y, const T* X, const T* label,
                                       const int N, const int D) {
  for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < N;
       i += blockDim.x * gridDim.x) {
    T sum = static_cast<T>(0);
    for (int j = 0; j < D; j++) {
42
      sum += label[i * D + j] * tolerable_value(log(X[i * D + j]));
43
    }
44
    Y[i] = -sum;
45 46 47
  }
}

48
// TODO(qingqing): make zero setting an common function.
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
template <typename T>
__global__ void zero(T* X, const int N) {
  for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < N;
       i += blockDim.x * gridDim.x) {
    X[i] = 0.0;
  }
}

template <typename T>
__global__ void CrossEntropyGradientKernel(T* dX, const T* dY, const T* X,
                                           const int* label, const int N,
                                           const int D) {
  // TOOD(qingqing) define CUDA_1D_KERNEL_LOOP macro in a common file.
  // CUDA_1D_KERNEL_LOOP(i, N) {
  for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < N;
       i += blockDim.x * gridDim.x) {
    int idx = i * D + label[i];
    dX[idx] = -dY[i] / X[idx];
  }
}

template <typename T>
71 72 73
__global__ void SoftCrossEntropyGradientKernel(T* dX, const T* dY, const T* X,
                                               const T* label, const int N,
                                               const int D) {
74
  // TOOD(qingqing): optimize for this kernel
75 76 77 78 79 80 81 82 83 84 85
  for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < N;
       i += blockDim.x * gridDim.x) {
    for (int j = 0; j < D; ++j) {
      int idx = i * D + j;
      dX[idx] = -label[idx] * dY[i] / X[idx];
    }
  }
}

template <typename T>
class CrossEntropyOpCUDAKernel : public framework::OpKernel {
86 87 88 89 90
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use GPUPlace.");

91 92 93 94 95 96 97
    auto x = ctx.Input<Tensor>("X");
    auto y = ctx.Output<Tensor>("Y");
    auto label = ctx.Input<Tensor>("Label");

    auto* x_data = x->data<T>();
    y->mutable_data<T>(ctx.GetPlace());
    auto* y_data = y->data<T>();
98

99 100
    int n = x->dims()[0];
    int d = x->dims()[1];
101
    int block = 512;
102
    int grid = (n + block - 1) / block;
103 104
    // TODO(qingqing) launch kernel on specified stream
    // base on ExecutionContext.
105
    if (ctx.Attr<bool>("soft_label")) {
106 107 108 109 110 111 112
      auto* label_data = ctx.Input<Tensor>("Label")->data<T>();
      SoftCrossEntropyKernel<T><<<grid, block>>>(y_data, x_data, label_data, n,
                                                 d);
    } else {
      auto* label_data = ctx.Input<Tensor>("Label")->data<int>();
      CrossEntropyKernel<T><<<grid, block>>>(y_data, x_data, label_data, n, d);
    }
113 114 115 116
  }
};

template <typename T>
117
class CrossEntropyGradientOpCUDAKernel : public framework::OpKernel {
118 119 120 121 122
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
    PADDLE_ENFORCE(platform::is_gpu_place(ctx.GetPlace()),
                   "It must use GPUPlace.");

123 124 125 126
    auto x = ctx.Input<Tensor>("X");
    auto dx = ctx.Output<Tensor>(framework::GradVarName("X"));
    auto dy = ctx.Input<Tensor>(framework::GradVarName("Y"));
    auto label = ctx.Input<Tensor>("Label");
127

128 129 130
    auto* dx_data = dx->mutable_data<T>(ctx.GetPlace());
    auto* dy_data = dy->data<T>();
    auto* x_data = x->data<T>();
131

132 133
    int n = x->dims()[0];
    int d = x->dims()[1];
134
    int block = 512;
135 136 137
    int grid = (n * d + block - 1) / block;
    zero<T><<<grid, block>>>(dx_data, n * d);
    grid = (n + block - 1) / block;
138 139
    // TODO(qingqing): launch kernel on specified stream
    // base on ExecutionContext.
140
    if (ctx.Attr<bool>("soft_label")) {
141 142 143 144 145 146 147 148
      auto* label_data = label->data<T>();
      SoftCrossEntropyGradientKernel<T><<<grid, block>>>(
          dx_data, dy_data, x_data, label_data, n, d);
    } else {
      auto* label_data = label->data<int>();
      CrossEntropyGradientKernel<T><<<grid, block>>>(dx_data, dy_data, x_data,
                                                     label_data, n, d);
    }
149 150 151 152 153
  }
};

}  // namespace operators
}  // namespace paddle
Q
Qiao Longfei 已提交
154

D
dongzhihong 已提交
155
namespace ops = paddle::operators;
156 157 158
REGISTER_OP_GPU_KERNEL(cross_entropy, ops::CrossEntropyOpCUDAKernel<float>);
REGISTER_OP_GPU_KERNEL(cross_entropy_grad,
                       ops::CrossEntropyGradientOpCUDAKernel<float>);