Merge pull request #503 from cocodark/develop

optimize pool 3x3s1p1

doc/design_doc.md

@@ -3,7 +3,6 @@
 
 #### 以下是 paddle-mobile 代码的执行流程图:
 
-
 ![执行流程图](http://otkwwi4x8.bkt.clouddn.com/2018-07-02-15305189473720.png)
 
 
@@ -15,7 +14,6 @@
 先来看一下模型, 模型分为两种结构:
  一种为参数文件是散开的, 如下图, 红框为模型结构的 protobuf 文件, 其余为参数文件
 
-
 ![模型描述](http://otkwwi4x8.bkt.clouddn.com/2018-07-02-15305190629577.png)
 
 
@@ -23,6 +21,7 @@
 
 ![模型描述combined](http://otkwwi4x8.bkt.clouddn.com/2018-07-02-15305191057130.png)
 
+
 loader 模块的作用是将模型结构信息 load 进内存, 将红框内的 protobuf 文件 load 进内存, 并对模型结构进行优化(如将几个细粒度的 op 融合成 粗粒度的 op, 如将 conv、 add、 batchnorm、 relu 融合为 conv\_add\_batchnorm\_relu).
 方便进行算法优化.
 

doc/development_doc.md

+### iOS&Android开发文档
+
 # iOS开发文档
 
 ## 编译
@@ -72,14 +74,154 @@ PaddleMobile.h
 ```
 
 
+#Android开发文档
+用户可通过如下两种方式，交叉编译Android平台上适用的paddle-mobile库：
+
+- 基于Docker容器编译
+- 基于Linux交叉编译
+
+
+## 基于Docker容器编译
+### 1. 安装 docker
+安装 docker 的方式，参考官方文档 [https://docs.docker.com/install/](https://docs.docker.com/install/)
+### 2. 使用 docker 搭建构建环境
+首先进入 paddle-mobile 的目录下，执行 `docker build`
+以 Linux/Mac 为例 (windows 建议在 'Docker Quickstart Terminal' 中执行)
+
+```
+$ docker build -t paddle-mobile:dev - < Dockerfile
+```
+使用 `docker images` 可以看到我们新建的 image
+
+```
+$ docker images
+REPOSITORY      TAG     IMAGE ID       CREATED         SIZE
+paddle-mobile   dev     33b146787711   45 hours ago    372MB
+```
+### 3. 使用 docker 构建
+进入 paddle-mobile 目录，执行 docker run
+
+```
+$ docker run -it --mount type=bind,source=$PWD,target=/paddle-mobile paddle-mobile:dev
+root@5affd29d4fc5:/ # cd /paddle-mobile
+# 生成构建 android 产出的 Makefile
+root@5affd29d4fc5:/ # rm CMakeCache.txt
+root@5affd29d4fc5:/ # cmake -DCMAKE_TOOLCHAIN_FILE=tools/toolchains/arm-android-neon.cmake
+# 生成构建 linux 产出的 Makefile
+root@5affd29d4fc5:/ # rm CMakeCache.txt
+root@5affd29d4fc5:/ # cmake -DCMAKE_TOOLCHAIN_FILE=tools/toolchains/arm-linux-gnueabi.cmake
+```
+### 4. 设置编译选项
+可以通过 ccmake 设置编译选项
+
+```
+root@5affd29d4fc5:/ # ccmake .
+                                                     Page 1 of 1
+ CMAKE_ASM_FLAGS
+ CMAKE_ASM_FLAGS_DEBUG
+ CMAKE_ASM_FLAGS_RELEASE
+ CMAKE_BUILD_TYPE
+ CMAKE_INSTALL_PREFIX             /usr/local
+ CMAKE_TOOLCHAIN_FILE             /paddle-mobile/tools/toolchains/arm-android-neon.cmake
+ CPU                              ON
+ DEBUGING                         ON
+ FPGA                             OFF
+ LOG_PROFILE                      ON
+ MALI_GPU                         OFF
+ NET                              googlenet
+ USE_EXCEPTION                    ON
+ USE_OPENMP                       OFF
+```
+修改选项后，按 `c`, `g` 更新 Makefile
+### 5. 构建
+使用 make 命令进行构建
+
+```
+root@5affd29d4fc5:/ # make
+```
+### 6. 查看构建产出
+构架产出可以在 host 机器上查看，在 paddle-mobile 的目录下，build 以及 test/build 下，可以使用 adb 指令或者 scp 传输到 device 上执行
+
+## 基于Linux交叉编译
+### 交叉编译环境准备
+##### 下载Android NDK
+
+从源码交叉编译paddle-mobile,用户需要提前准备好交叉编译环境。Android平台使用的C/C++交叉编译工具链是[Android NDK](https://developer.android.com/ndk/)，用户可以自行前往下载，也可以通过以下命令获取：
+
+```
+wget https://dl.google.com/android/repository/android-ndk-r17b-darwin-x86_64.zip
+unzip android-ndk-r17b-darwin-x86_64.zip
+
+```
+
+##### 设置环境变量
+工程中自带的独立工具链会根据环境变量NDK_ROOT查找NDK，因此需要配置环境变量：
+
+```
+export NDK_ROOT = "path to ndk"
+```
+### 执行编译
+在paddle-mobile根目录中，执行以下命令：
+
+```
+cd tools
+sh build.sh android
+
+```
+执行完毕后，生成的so位于build目录中，单测可执行文件位于test/build目录中。
+##### Tips:
+如果想要获得体积更小的库，可选择编译支持指定模型结构的库。
+如执行如下命令：
+
+```
+sh build.sh android googlenet
+```
+会得到一个支持googlnet的体积更小的库。
 
+##测试
+在编译完成后，我们提供了自动化的测试脚本，帮助用户将运行单测文件所需要的模型及库文件push到Android设备中，执行以下命令：
 
+```
+cd tools/android-debug-script
+sh run_on_android.sh (npm) 可选参数npm,用于选择是否传输模型文件到手机上
+```
+出现如下提示：
 
+```
+**** choose OP or NET to test ****
+which to test :
+```
+输入名称即可运行对应的测试文件。
 
+##部署
+Android应用可通过JNI接口调用底层C/C++，paddle-mobile对外提供的JNI接口如下：
 
+##### 1 load接口 加载模型参数
 
+```
+/*
+*@param modelPath 模型文件路径
+*@return jboolean
+*/
+JNIEXPORT jboolean JNICALL Java_com_baidu_paddle_PML_load(JNIEnv *env,
+                                                          jclass thiz,
+                                                          jstring modelPath);
+```
 
+##### 2 predict接口 执行预测
 
+```
+/**
+*@param buf 输入数据
+*@return 输出数据
+JNIEXPORT jfloatArray JNICALL Java_com_baidu_paddle_PML_predict(
+    JNIEnv *env, jclass thiz, jfloatArray buf);
+```
+##### 3 clear接口 销毁实例、清理内存操作
 
+```
+JNIEXPORT void JNICALL Java_com_baidu_paddle_PMLL_clear(JNIEnv *env,
+                                                        jclass thiz);
+```
 
 

src/framework/scope.h

@@ -23,7 +23,17 @@ namespace framework {
 class Scope {
  public:
   Scope() = default;
-  ~Scope() = default;
+
+  ~Scope() {
+    for (auto &var : vars_) {
+      delete var.second;
+    }
+    vars_.clear();
+    for (auto kid : kids_) {
+      delete kid;
+    }
+    kids_.clear();
+  }
 
   Scope &NewScope() const;
 

src/jni/paddle_mobile_jni.cpp

@@ -54,13 +54,14 @@ string jstring2cppstring(JNIEnv *env, jstring jstr) {
 JNIEXPORT jboolean JNICALL Java_com_baidu_paddle_PML_load(JNIEnv *env,
                                                           jclass thiz,
                                                           jstring modelPath) {
+  ANDROIDLOGI("load invoked");
   bool optimize = true;
   return getPaddleMobileInstance()->Load(jstring2cppstring(env, modelPath),
                                          optimize);
 }
 
-JNIEXPORT jfloatArray JNICALL Java_com_baidu_paddle_PML_predictImage(
-    JNIEnv *env, jclass thiz, jfloatArray buf) {
+JNIEXPORT jfloatArray JNICALL
+Java_com_baidu_paddle_PML_predict(JNIEnv *env, jclass thiz, jfloatArray buf) {
   jfloatArray result = NULL;
   int count = 0;
   float *dataPointer = nullptr;
@@ -78,6 +79,7 @@ JNIEXPORT jfloatArray JNICALL Java_com_baidu_paddle_PML_predictImage(
   count = output->numel();
   result = env->NewFloatArray(count);
   env->SetFloatArrayRegion(result, 0, count, output->data<float>());
+  ANDROIDLOGI("predict finished");
   return result;
 }
 

src/jni/paddle_mobile_jni.h

@@ -31,8 +31,8 @@ JNIEXPORT jboolean JNICALL Java_com_baidu_paddle_PML_load(JNIEnv *env,
 /**
  * object detection for anroid
  */
-JNIEXPORT jfloatArray JNICALL Java_com_baidu_paddle_PML_predictImage(
-    JNIEnv *env, jclass thiz, jfloatArray buf);
+JNIEXPORT jfloatArray JNICALL
+Java_com_baidu_paddle_PML_predict(JNIEnv *env, jclass thiz, jfloatArray buf);
 
 /**
  * clear data of the net when destroy for android

src/operators/kernel/arm/conv_add_kernel.cpp

@@ -14,6 +14,7 @@ limitations under the License. */
 #ifdef FUSION_CONVADD_OP
 
 #include "operators/kernel/conv_add_kernel.h"
+#include "../central-arm-func/conv_add_arm_func.h"
 
 namespace paddle_mobile {
 namespace operators {
@@ -23,111 +24,9 @@ bool ConvAddKernel<CPU, float>::Init(FusionConvAddParam *param) {
   return true;
 }
 
-void ConvAddBasic(const FusionConvAddParam &param) {
-  const Tensor *input = param.Input();
-  Tensor filter = *param.Filter();
-  Tensor bias = *param.Bias();
-  int axis = param.Axis();
-  Tensor *output = param.Output();
-  math::expand_bias(bias, axis, output->dims());
-  output->ShareDataWith(bias);
-  int groups = param.Groups();
-  std::vector<int> strides = param.Strides();
-  std::vector<int> paddings = param.Paddings();
-  std::vector<int> dilations = param.Dilations();
-
-  const int batch_size = static_cast<int>(input->dims()[0]);
-
-  std::vector<int64_t> filter_shape_vec(framework::vectorize(filter.dims()));
-
-  std::vector<int64_t> output_shape_vec(framework::vectorize(output->dims()));
-  size_t data_dim = filter_shape_vec.size() - 2;
-  std::vector<int64_t> col_shape_vec(1 + 2 * data_dim);
-  col_shape_vec[0] = input->dims()[1] / groups;
-  for (size_t j = 0; j < data_dim; ++j) {
-    col_shape_vec[j + 1] = filter_shape_vec[j + 2];
-    col_shape_vec[j + 1 + data_dim] = output_shape_vec[j + 2];
-  }
-  framework::DDim col_shape(framework::make_ddim(col_shape_vec));
-
-  framework::DDim col_matrix_shape =
-      framework::flatten_to_2d(col_shape, data_dim + 1);
-
-  bool is_expand =
-      math::IsExpand(filter_shape_vec, strides, paddings, dilations);
-  Tensor col;
-  Tensor col_matrix;
-  if (is_expand) {
-    col.mutable_data<float>(col_shape);
-    col_matrix.ShareDataWith(col);
-    col_matrix.Resize(col_matrix_shape);
-  }
-
-  framework::DDim input_shape = framework::slice_ddim(
-      input->dims(), 1, static_cast<int>(input->dims().size()));
-
-  framework::DDim filter_matrix_shape = {filter.dims()[0],
-                                         filter.numel() / filter.dims()[0]};
-  filter.Resize(filter_matrix_shape);
-  framework::DDim output_matrix_shape = {
-      output->dims()[1],
-      output->numel() / (output->dims()[0] * output->dims()[1])};
-
-  // convolution operator: im2col(or vol2col) + gemm
-  int in_step = static_cast<int>(input->dims()[1]) / groups;
-  int out_step = static_cast<int>(output->dims()[1]) / groups;
-
-  math::Vol2ColFunctor<CPU, float> vol2col;
-  math::Im2ColFunctor<math::ColFormat::kCFO, CPU, float> im2col;
-
-  for (int i = 0; i < batch_size; i++) {
-    Tensor in_batch = input->Slice(i, i + 1).Resize(input_shape);
-    Tensor out_batch = output->Slice(i, i + 1).Resize(output_matrix_shape);
-
-    for (int g = 0; g < groups; g++) {
-      Tensor in_slice = in_batch.Slice(g * in_step, (g + 1) * in_step);
-
-      if (!is_expand) {
-        col.ShareDataWith(in_slice);
-        col_matrix.ShareDataWith(col);
-        col_matrix.Resize(col_matrix_shape);
-      } else if (data_dim == 2U) {
-        // im2col
-        im2col(in_slice, dilations, strides,
-               std::vector<int>{paddings[0], paddings[1], paddings[0],
-                                paddings[1]},
-               &col);
-      } else if (data_dim == 3U) {
-        // vol2col
-        vol2col(in_slice, dilations, strides, paddings, &col);
-      }
-      // gemm
-      Tensor out_slice = out_batch.Slice(g * out_step, (g + 1) * out_step);
-      Tensor filter_slice = filter.Slice(g * out_step, (g + 1) * out_step);
-      math::matmul<float>(filter_slice, false, col_matrix, false,
-                          static_cast<float>(1), &out_slice,
-                          static_cast<float>(1));
-    }
-  }
-}
-
 template <>
 void ConvAddKernel<CPU, float>::Compute(const FusionConvAddParam &param) const {
-  if (param.Groups() == param.Input()->dims()[1] &&
-      param.Input()->dims()[1] == param.Output()->dims()[1] &&
-      param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
-      param.Filter()->dims()[2] == 3 && param.Strides()[0] == 1) {
-    math::DepthwiseConv3x3s1p1(param.Input(), param.Filter(), param.Output(),
-                               param.Bias(), true);
-  } else if (param.Groups() == param.Input()->dims()[1] &&
-             param.Input()->dims()[1] == param.Output()->dims()[1] &&
-             param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
-             param.Filter()->dims()[2] == 3) {
-    math::DepthwiseConv3x3(param.Input(), param.Strides(), param.Paddings(),
-                           param.Filter(), param.Bias(), param.Output(), true);
-  } else {
-    ConvAddBasic(param);
-  }
+  ConvAddCompute<float>(param);
 }
 
 template class ConvAddKernel<CPU, float>;

src/operators/kernel/arm/pool_kernel.cpp

@@ -14,27 +14,11 @@ limitations under the License. */
 
 #ifdef POOL_OP
 
-#include <operators/kernel/pool_kernel.h>
-#include "common/log.h"
-
+#include "operators/kernel/pool_kernel.h"
+#include "../central-arm-func/pool_arm_func.h"
 namespace paddle_mobile {
 namespace operators {
 
-inline void PoolBasic(std::string pooling_type, std::vector<int> ksize,
-                      std::vector<int> strides, std::vector<int> paddings,
-                      const Tensor *in_x, Tensor *out) {
-  if (pooling_type == "max") {
-    math::PoolFunctor<CPU, math::MaxPool<float>, float> pool2d_forward;
-    math::MaxPool<float> pool_process;
-    pool2d_forward(*in_x, ksize, strides, paddings, pool_process, out);
-
-  } else if (pooling_type == "avg") {
-    math::PoolFunctor<CPU, math::AvgPool<float>, float> pool2d_forward;
-    math::AvgPool<float> pool_process;
-    pool2d_forward(*in_x, ksize, strides, paddings, pool_process, out);
-  }
-}
-
 template <>
 bool PoolKernel<CPU, float>::Init(PoolParam *param) {
   return true;
@@ -42,42 +26,7 @@ bool PoolKernel<CPU, float>::Init(PoolParam *param) {
 
 template <>
 void PoolKernel<CPU, float>::Compute(const PoolParam &param) const {
-  const Tensor *in_x = param.Input();
-  Tensor *out = param.Output();
-  std::string pooling_type = param.PoolingType();
-
-  std::vector<int> ksize = param.Ksize();
-
-  std::vector<int> strides = param.Strides();
-
-  std::vector<int> paddings = param.Paddings();
-  if (ksize.size() != 2) {
-    LOG(paddle_mobile::LogLevel::kLOG_ERROR)
-        << "Pool op only supports 2D and 3D input.";
-  }
-
-  if (param.isGlobalPooling()) {
-    for (size_t i = 0; i < ksize.size(); ++i) {
-      paddings[i] = 0;
-      ksize[i] = static_cast<int>(in_x->dims()[i + 2]);
-    }
-  } else if (ksize[0] == 3 && ksize[0] == ksize[1]) {
-    if (pooling_type == "max") {
-      math::Pool3x3Max(strides, paddings, in_x, out);
-    } else if (pooling_type == "avg") {
-      math::Pool3x3Avg(strides, paddings, in_x, out);
-    }
-
-  } else if (ksize[0] == 2 && ksize[0] == ksize[1]) {
-    if (pooling_type == "max") {
-      math::Pool2x2Max(strides, paddings, in_x, out);
-    } else if (pooling_type == "avg") {
-      math::Pool2x2Avg(strides, paddings, in_x, out);
-    }
-
-  } else {
-    PoolBasic(pooling_type, ksize, strides, paddings, in_x, out);
-  }
+  PoolCompute<float>(param);
 }
 }  // namespace operators
 }  // namespace paddle_mobile

src/operators/kernel/arm/sigmoid_kernel.cpp

@@ -15,6 +15,7 @@ limitations under the License. */
 #ifdef SIGMOID_OP
 
 #include "../sigmoid_kernel.h"
+#include "../central-arm-func/sigmoid_arm_func.h"
 #if __ARM_NEON
 #include "../../math/math_func_neon.h"
 #endif
@@ -25,52 +26,6 @@ namespace operators {
 using framework::DDim;
 using framework::Tensor;
 
-void sigmoid(const Tensor *X, Tensor *Y) {
-#if __ARM_NEON
-  const float *input = X->data<float>();
-  float *output = Y->mutable_data<float>();
-  const DDim &dDim = X->dims();
-  int axis_index = 1;
-  if (dDim.size() < 4) {
-    axis_index = 0;
-  }
-  DDim outer_ddim =
-      paddle_mobile::framework::slice_ddim(dDim, 0, axis_index + 1);
-  DDim inner_ddim =
-      paddle_mobile::framework::slice_ddim(dDim, axis_index + 1, dDim.size());
-  int out_size = paddle_mobile::framework::product(outer_ddim);
-  int inner_size = paddle_mobile::framework::product(inner_ddim);
-
-  DLOG << "outsize=" << out_size;
-  DLOG << "innersize=" << inner_size;
-  #pragma omp parallel for
-  for (int i = 0; i < out_size; ++i) {
-    const float *input_outer_ptr = input + i * inner_size;
-    float *output_outer_ptr = output + i * inner_size;
-    int nn = inner_size >> 2;
-    int remain = inner_size - (nn << 2);
-    float32x4_t _one = vdupq_n_f32(1.f);
-    for (; nn > 0; nn--) {
-      float32x4_t data = vld1q_f32(input_outer_ptr);
-      data = vnegq_f32(data);
-      data = exp_ps(data);
-      data = vaddq_f32(data, _one);
-      float32x4_t out_data = vrecpeq_f32(data);
-      out_data = vmulq_f32(vrecpsq_f32(data, out_data), out_data);
-      vst1q_f32(output_outer_ptr, out_data);
-
-      input_outer_ptr += 4;
-      output_outer_ptr += 4;
-    }
-    for (; remain > 0; remain--) {
-      *output_outer_ptr = 1.f / (1.f + exp(-*input_outer_ptr));
-      output_outer_ptr++;
-      input_outer_ptr++;
-    }
-  }
-#endif
-}
-
 template <>
 bool SigmoidKernel<CPU, float>::Init(SigmoidParam *param) {
   return true;
@@ -78,11 +33,7 @@ bool SigmoidKernel<CPU, float>::Init(SigmoidParam *param) {
 
 template <>
 void SigmoidKernel<CPU, float>::Compute(const SigmoidParam &param) const {
-  const Tensor *in_x = param.InputX();
-  Tensor *out = param.Out();
-  auto x_dims = in_x->dims();
-  out->Resize(x_dims);
-  sigmoid(in_x, out);
+  SigmoidCompute<float>(param);
 }
 
 template class SigmoidKernel<CPU, float>;

src/operators/kernel/arm/softmax_kernel.cpp

@@ -15,7 +15,8 @@ limitations under the License. */
 #ifdef SOFTMAX_OP
 
 #include "../softmax_kernel.h"
-#include "../../math/softmax.h"
+#include "../central-arm-func/softmax_arm_func.h"
+#include "operators/math/softmax.h"
 namespace paddle_mobile {
 namespace operators {
 
@@ -26,11 +27,7 @@ bool SoftmaxKernel<CPU, float>::Init(SoftmaxParam *param) {
 
 template <>
 void SoftmaxKernel<CPU, float>::Compute(const SoftmaxParam &param) const {
-  const Tensor *in_x = param.InputX();
-  Tensor *out = param.Out();
-  auto x_dims = in_x->dims();
-  out->Resize(x_dims);
-  math::SoftmaxFuntor<CPU, float>()(in_x, out);
+  SoftmaxCompute<float>(param);
 }
 
 template class SoftmaxKernel<CPU, float>;

src/operators/kernel/central-arm-func/conv_add_arm_func.h

+/* 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. */
+
+#ifdef FUSION_CONVADD_OP
+#pragma once
+
+#include <vector>
+#include "operators/math/conv_func.h"
+#include "operators/math/depthwise_conv_3x3.h"
+#include "operators/math/im2col.h"
+#include "operators/math/math_function.h"
+#include "operators/math/vol2col.h"
+#include "operators/op_param.h"
+
+namespace paddle_mobile {
+namespace operators {
+void ConvAddBasic(const FusionConvAddParam &param) {
+  const Tensor *input = param.Input();
+  Tensor filter = *param.Filter();
+  Tensor bias = *param.Bias();
+  int axis = param.Axis();
+  Tensor *output = param.Output();
+  math::expand_bias(bias, axis, output->dims());
+  output->ShareDataWith(bias);
+  int groups = param.Groups();
+  std::vector<int> strides = param.Strides();
+  std::vector<int> paddings = param.Paddings();
+  std::vector<int> dilations = param.Dilations();
+
+  const int batch_size = static_cast<int>(input->dims()[0]);
+
+  std::vector<int64_t> filter_shape_vec(framework::vectorize(filter.dims()));
+
+  std::vector<int64_t> output_shape_vec(framework::vectorize(output->dims()));
+  size_t data_dim = filter_shape_vec.size() - 2;
+  std::vector<int64_t> col_shape_vec(1 + 2 * data_dim);
+  col_shape_vec[0] = input->dims()[1] / groups;
+  for (size_t j = 0; j < data_dim; ++j) {
+    col_shape_vec[j + 1] = filter_shape_vec[j + 2];
+    col_shape_vec[j + 1 + data_dim] = output_shape_vec[j + 2];
+  }
+  framework::DDim col_shape(framework::make_ddim(col_shape_vec));
+
+  framework::DDim col_matrix_shape =
+      framework::flatten_to_2d(col_shape, data_dim + 1);
+
+  bool is_expand =
+      math::IsExpand(filter_shape_vec, strides, paddings, dilations);
+  Tensor col;
+  Tensor col_matrix;
+  if (is_expand) {
+    col.mutable_data<float>(col_shape);
+    col_matrix.ShareDataWith(col);
+    col_matrix.Resize(col_matrix_shape);
+  }
+
+  framework::DDim input_shape = framework::slice_ddim(
+      input->dims(), 1, static_cast<int>(input->dims().size()));
+
+  framework::DDim filter_matrix_shape = {filter.dims()[0],
+                                         filter.numel() / filter.dims()[0]};
+  filter.Resize(filter_matrix_shape);
+  framework::DDim output_matrix_shape = {
+      output->dims()[1],
+      output->numel() / (output->dims()[0] * output->dims()[1])};
+
+  // convolution operator: im2col(or vol2col) + gemm
+  int in_step = static_cast<int>(input->dims()[1]) / groups;
+  int out_step = static_cast<int>(output->dims()[1]) / groups;
+
+  math::Vol2ColFunctor<CPU, float> vol2col;
+  math::Im2ColFunctor<math::ColFormat::kCFO, CPU, float> im2col;
+
+  for (int i = 0; i < batch_size; i++) {
+    Tensor in_batch = input->Slice(i, i + 1).Resize(input_shape);
+    Tensor out_batch = output->Slice(i, i + 1).Resize(output_matrix_shape);
+
+    for (int g = 0; g < groups; g++) {
+      Tensor in_slice = in_batch.Slice(g * in_step, (g + 1) * in_step);
+
+      if (!is_expand) {
+        col.ShareDataWith(in_slice);
+        col_matrix.ShareDataWith(col);
+        col_matrix.Resize(col_matrix_shape);
+      } else if (data_dim == 2U) {
+        // im2col
+        im2col(in_slice, dilations, strides,
+               std::vector<int>{paddings[0], paddings[1], paddings[0],
+                                paddings[1]},
+               &col);
+      } else if (data_dim == 3U) {
+        // vol2col
+        vol2col(in_slice, dilations, strides, paddings, &col);
+      }
+      // gemm
+      Tensor out_slice = out_batch.Slice(g * out_step, (g + 1) * out_step);
+      Tensor filter_slice = filter.Slice(g * out_step, (g + 1) * out_step);
+      math::matmul<float>(filter_slice, false, col_matrix, false,
+                          static_cast<float>(1), &out_slice,
+                          static_cast<float>(1));
+    }
+  }
+}
+
+template <typename P>
+void ConvAddCompute(const FusionConvAddParam &param) {
+  if (param.Groups() == param.Input()->dims()[1] &&
+      param.Input()->dims()[1] == param.Output()->dims()[1] &&
+      param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
+      param.Filter()->dims()[2] == 3 && param.Strides()[0] == 1) {
+    math::DepthwiseConv3x3s1p1(param.Input(), param.Filter(), param.Output(),
+                               param.Bias(), true);
+  } else if (param.Groups() == param.Input()->dims()[1] &&
+             param.Input()->dims()[1] == param.Output()->dims()[1] &&
+             param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
+             param.Filter()->dims()[2] == 3) {
+    math::DepthwiseConv3x3(param.Input(), param.Strides(), param.Paddings(),
+                           param.Filter(), param.Bias(), param.Output(), true);
+  } else {
+    ConvAddBasic(param);
+  }
+}
+
+}  // namespace operators
+}  // namespace paddle_mobile
+
+#endif

src/operators/kernel/central-arm-func/conv_arm_func.h

@@ -15,19 +15,21 @@ limitations under the License. */
 #ifdef CONV_OP
 
 #pragma once
-#include <operators/math/depthwise_conv_3x3.h>
 #include <vector>
-
+#include "operators/math/conv_func.h"
+#include "operators/math/depthwise_conv_3x3.h"
+#include "operators/math/im2col.h"
+#include "operators/math/math_function.h"
+#include "operators/math/vol2col.h"
 #include "operators/op_param.h"
 
 namespace paddle_mobile {
 namespace operators {
-
 inline void ConvBasic(const ConvParam &param) {
   const Tensor *input = param.Input();
   Tensor filter = *param.Filter();
   Tensor *output = param.Output();
-
+  output->mutable_data<float>();
   int groups = param.Groups();
   std::vector<int> strides = param.Strides();
   std::vector<int> paddings = param.Paddings();
@@ -111,20 +113,18 @@ inline void ConvBasic(const ConvParam &param) {
 
 template <typename P>
 void ConvCompute(const ConvParam &param) {
-  Tensor Bias;
-  Bias.mutable_data<float>({param.Groups()});
   if (param.Groups() == param.Input()->dims()[1] &&
       param.Input()->dims()[1] == param.Output()->dims()[1] &&
       param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
       param.Filter()->dims()[2] == 3 && param.Strides()[0] == 1) {
     math::DepthwiseConv3x3s1p1(param.Input(), param.Filter(), param.Output(),
-                               &Bias, false);
+                               nullptr, false);
   } else if (param.Groups() == param.Input()->dims()[1] &&
              param.Input()->dims()[1] == param.Output()->dims()[1] &&
              param.Filter()->dims()[2] == param.Filter()->dims()[3] &&
-             param.Filter()->dims()[2] == 3 && param.Strides()[0] == 2) {
+             param.Filter()->dims()[2] == 3) {
     math::DepthwiseConv3x3(param.Input(), param.Strides(), param.Paddings(),
-                           param.Filter(), &Bias, param.Output(), false);
+                           param.Filter(), nullptr, param.Output(), false);
   } else {
     ConvBasic(param);
   }

src/operators/kernel/central-arm-func/pool_arm_func.h

+/* 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. */
+
+#ifdef POOL_OP
+#pragma once
+
+#include <string>
+#include <vector>
+#include "operators/math/pooling.h"
+
+namespace paddle_mobile {
+namespace operators {
+using framework::Tensor;
+
+inline void PoolBasic(std::string pooling_type, std::vector<int> ksize,
+                      std::vector<int> strides, std::vector<int> paddings,
+                      const Tensor *in_x, Tensor *out) {
+  if (pooling_type == "max") {
+    math::PoolFunctor<CPU, math::MaxPool<float>, float> pool2d_forward;
+    math::MaxPool<float> pool_process;
+    pool2d_forward(*in_x, ksize, strides, paddings, pool_process, out);
+
+  } else if (pooling_type == "avg") {
+    math::PoolFunctor<CPU, math::AvgPool<float>, float> pool2d_forward;
+    math::AvgPool<float> pool_process;
+    pool2d_forward(*in_x, ksize, strides, paddings, pool_process, out);
+  }
+}
+template <typename P>
+void PoolCompute(const PoolParam &param) {
+  const Tensor *in_x = param.Input();
+  Tensor *out = param.Output();
+  std::string pooling_type = param.PoolingType();
+
+  std::vector<int> ksize = param.Ksize();
+
+  std::vector<int> strides = param.Strides();
+
+  std::vector<int> paddings = param.Paddings();
+  if (ksize.size() != 2) {
+    LOG(paddle_mobile::LogLevel::kLOG_ERROR)
+        << "Pool op only supports 2D and 3D input.";
+  }
+
+  if (param.isGlobalPooling()) {
+    for (size_t i = 0; i < ksize.size(); ++i) {
+      paddings[i] = 0;
+      ksize[i] = static_cast<int>(in_x->dims()[i + 2]);
+    }
+  } else if (ksize[0] == 3 && ksize[0] == ksize[1]) {
+    if (pooling_type == "max") {
+      if (strides[0] == strides[1] && strides[0] == 1 &&
+          paddings[0] == paddings[1] && paddings[1] == 1) {
+        math::Pool3x3Maxs1p1(in_x, out);
+      } else {
+        math::Pool3x3Max(strides, paddings, in_x, out);
+      }
+    } else if (pooling_type == "avg") {
+      if (strides[0] == strides[1] && strides[0] == 1 &&
+          paddings[0] == paddings[1] && paddings[1] == 1) {
+        math::Pool3x3Avgs1p1(in_x, out);
+      } else {
+        math::Pool3x3Avg(strides, paddings, in_x, out);
+      }
+    }
+
+  } else if (ksize[0] == 2 && ksize[0] == ksize[1]) {
+    if (pooling_type == "max") {
+      math::Pool2x2Max(strides, paddings, in_x, out);
+    } else if (pooling_type == "avg") {
+      math::Pool2x2Avg(strides, paddings, in_x, out);
+    }
+
+  } else {
+    PoolBasic(pooling_type, ksize, strides, paddings, in_x, out);
+  }
+}
+
+}  // namespace operators
+}  // namespace paddle_mobile
+#endif

src/operators/kernel/central-arm-func/sigmoid_arm_func.h

+/* 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. */
+#ifdef SIGMOID_OP
+#pragma once
+
+#include "operators/op_param.h"
+#if __ARM_NEON
+#include <arm_neon.h>
+#include "operators/math/math_func_neon.h"
+#endif
+
+namespace paddle_mobile {
+namespace operators {
+using framework::DDim;
+void sigmoid(const Tensor *X, Tensor *Y) {
+#if __ARM_NEON
+  const float *input = X->data<float>();
+  float *output = Y->mutable_data<float>();
+  const DDim &dDim = X->dims();
+  int axis_index = 1;
+  if (dDim.size() < 4) {
+    axis_index = 0;
+  }
+  DDim outer_ddim =
+      paddle_mobile::framework::slice_ddim(dDim, 0, axis_index + 1);
+  DDim inner_ddim =
+      paddle_mobile::framework::slice_ddim(dDim, axis_index + 1, dDim.size());
+  int out_size = paddle_mobile::framework::product(outer_ddim);
+  int inner_size = paddle_mobile::framework::product(inner_ddim);
+
+  DLOG << "outsize=" << out_size;
+  DLOG << "innersize=" << inner_size;
+  #pragma omp parallel for
+  for (int i = 0; i < out_size; ++i) {
+    const float *input_outer_ptr = input + i * inner_size;
+    float *output_outer_ptr = output + i * inner_size;
+    int nn = inner_size >> 2;
+    int remain = inner_size - (nn << 2);
+    float32x4_t _one = vdupq_n_f32(1.f);
+    for (; nn > 0; nn--) {
+      float32x4_t data = vld1q_f32(input_outer_ptr);
+      data = vnegq_f32(data);
+      data = exp_ps(data);
+      data = vaddq_f32(data, _one);
+      float32x4_t out_data = vrecpeq_f32(data);
+      out_data = vmulq_f32(vrecpsq_f32(data, out_data), out_data);
+      vst1q_f32(output_outer_ptr, out_data);
+
+      input_outer_ptr += 4;
+      output_outer_ptr += 4;
+    }
+    for (; remain > 0; remain--) {
+      *output_outer_ptr = 1.f / (1.f + exp(-*input_outer_ptr));
+      output_outer_ptr++;
+      input_outer_ptr++;
+    }
+  }
+#endif
+}
+
+template <typename P>
+void SigmoidCompute(const SigmoidParam &param) {
+  const Tensor *in_x = param.InputX();
+  Tensor *out = param.Out();
+  auto x_dims = in_x->dims();
+  out->Resize(x_dims);
+  sigmoid(in_x, out);
+}
+}  // namespace operators
+}  // namespace paddle_mobile
+#endif

src/operators/kernel/central-arm-func/softmax_arm_func.h

+/* 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. */
+
+#ifdef SOFTMAX_OP
+#pragma once
+#include "../../math/softmax.h"
+namespace paddle_mobile {
+namespace operators {
+template <typename P>
+void SoftmaxCompute(const SoftmaxParam &param) {
+  const Tensor *in_x = param.InputX();
+  Tensor *out = param.Out();
+  auto x_dims = in_x->dims();
+  out->Resize(x_dims);
+  math::SoftmaxFuntor<CPU, float>()(in_x, out);
+}
+}  // namespace operators
+}  // namespace paddle_mobile
+#endif

src/operators/kernel/pool_kernel.h

@@ -17,7 +17,6 @@ limitations under the License. */
 #pragma once
 
 #include "framework/operator.h"
-#include "operators/math/pooling.h"
 #include "operators/op_param.h"
 
 namespace paddle_mobile {

src/operators/kernel/softmax_kernel.h

@@ -23,8 +23,6 @@ namespace paddle_mobile {
 namespace operators {
 using framework::OpKernelBase;
 
-void simoid(Tensor *X, Tensor *Y);
-
 template <typename DeviceType, typename T>
 class SoftmaxKernel : public OpKernelBase<DeviceType, SoftmaxParam> {
  public:

src/operators/math/depthwise_conv_3x3.cpp

@@ -245,7 +245,10 @@ void DepthwiseConv3x3s1p1(const Tensor *input, const Tensor *filter,
   const float *input_data = input->data<float>();
   const float *filter_data = filter->data<float>();
   float *output_data = output->data<float>();
-  const float *bias_data = bias->data<float>();
+  const float *bias_data;
+  if (if_bias) {
+    bias_data = bias->data<float>();
+  }
 
   const int h = static_cast<int>(input->dims()[2]);
   const int w = static_cast<int>(input->dims()[3]);

src/operators/math/pool_3x3.cpp

@@ -13,7 +13,6 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #ifdef POOL_OP
-#define __ARM_NEON true
 #include "pool_3x3.h"
 #include "framework/tensor.h"
 #if __ARM_NEON
@@ -27,6 +26,481 @@ using framework::Tensor;
 using std::max;
 using std::min;
 using std::vector;
+void Pool3x3Avgs1p1(const Tensor *input, Tensor *output) {
+#if __ARM_NEON
+  const int batch_size = input->dims()[0];
+
+  const int h_in = input->dims()[2];
+
+  const int w_in = input->dims()[3];
+
+  const int output_channels = output->dims()[1];
+
+  const int h_out = output->dims()[2];
+  const int w_out = output->dims()[3];
+  const int outputdata_channel_stride = h_out * w_out;
+  const int inputdata_channel_stride = h_in * w_in;
+  float *out_data = output->data<float>();
+  const float *input_data = input->data<float>();
+  const float coef = 1.0 / 9.0;
+  for (int k = 0; k < batch_size; ++k) {
+    for (int c = 0; c < output_channels; ++c) {
+      // four corner point
+      out_data[0] = (input_data[0] + input_data[1] + input_data[w_in] +
+                     input_data[w_in + 1]) *
+                    coef;
+      out_data[w_out - 1] =
+          (input_data[w_in - 2] + input_data[w_in - 1] +
+           input_data[w_in * 2 - 2] + input_data[2 * w_in - 1]) *
+          coef;
+      out_data[(h_out - 1) * w_out] =
+          (input_data[(h_in - 2) * w_in] + input_data[(h_in - 2) * w_in + 1] +
+           input_data[(h_in - 1) * w_in] + input_data[(h_in - 1) * w_in + 1]) *
+          coef;
+      out_data[h_out * w_out - 1] =
+          (input_data[h_in * w_in - 1] + input_data[h_in * w_in - 2] +
+           input_data[(h_in - 1) * w_in - 1] +
+           input_data[(h_in - 1) * w_in - 2]) *
+          coef;
+      // left side & right side
+      for (int i = 1; i < h_in - 1; ++i) {
+        out_data[i * w_out] =
+            (input_data[i * w_in - w_in] + input_data[i * w_in - w_in + 1] +
+             input_data[i * w_in] + input_data[i * w_in + 1] +
+             input_data[i * w_in + w_in] + input_data[i * w_in + w_in + 1]) *
+            coef;
+        out_data[i * w_out + w_out - 1] =
+            (input_data[i * w_in - w_in + w_in - 2] +
+             input_data[i * w_in - w_in + 1 + w_in - 2] +
+             input_data[i * w_in + w_in - 2] +
+             input_data[i * w_in + 1 + w_in - 2] +
+             input_data[i * w_in + w_in + w_in - 2] +
+             input_data[i * w_in + w_in + 1 + w_in - 2]) *
+            coef;
+      }
+      // top 1 row & bottom 1 row
+      const float *input_tmp = input_data;
+
+      float32x4_t in0, in1, in2, in3, in4, in5, in6, in7, tmp0, tmp1, tmp2,
+          tmp3, tmp4, tmp5, sum, out0;
+      float32x4_t v_coef = vdupq_n_f32(coef);
+      in0 = vld1q_f32(input_tmp);
+      in2 = vld1q_f32(input_tmp + w_in);
+      const float *input_tmp_end = input_tmp + (h_in - 2) * w_in;
+      in4 = vld1q_f32(input_tmp_end);
+      in6 = vld1q_f32(input_tmp_end + w_in);
+      int c_mid = w_out - 2;
+      auto output_ptr = out_data + 1;
+      for (; c_mid > 3; c_mid -= 4) {
+        in1 = vld1q_f32(input_tmp + 4);
+        in3 = vld1q_f32(input_tmp + w_in + 4);
+
+        tmp0 = vextq_f32(in0, in1, 1);
+        tmp1 = vextq_f32(in0, in1, 2);
+
+        tmp2 = vextq_f32(in2, in3, 1);
+        tmp3 = vextq_f32(in2, in3, 2);
+
+        sum = vaddq_f32(in0, tmp0);
+        sum = vaddq_f32(sum, tmp1);
+        sum = vaddq_f32(sum, in2);
+        sum = vaddq_f32(sum, tmp2);
+        sum = vaddq_f32(sum, tmp3);
+
+        vst1q_f32(output_ptr, vmulq_f32(sum, v_coef));
+
+        in5 = vld1q_f32(input_tmp_end + 4);
+        in7 = vld1q_f32(input_tmp_end + w_in + 4);
+
+        tmp0 = vextq_f32(in4, in5, 1);
+        tmp1 = vextq_f32(in4, in5, 2);
+        tmp2 = vextq_f32(in6, in7, 1);
+        tmp3 = vextq_f32(in6, in7, 2);
+
+        sum = vaddq_f32(in0, tmp0);
+        sum = vaddq_f32(sum, tmp1);
+        sum = vaddq_f32(sum, in2);
+        sum = vaddq_f32(sum, tmp2);
+        sum = vaddq_f32(sum, tmp3);
+
+        vst1q_f32(output_ptr + (h_out - 1) * w_out, vmulq_f32(sum, v_coef));
+
+        // can optimize to each 8 stride.
+        input_tmp += 4;
+        input_tmp_end += 4;
+        output_ptr += 4;
+        in0 = in1;
+        in2 = in3;
+        in4 = in5;
+        in6 = in7;
+      }
+      // top right remain
+      float32x4_t pad0 = vdupq_n_f32(input_data[w_in - 1]);
+      float32x4_t pad1 = vdupq_n_f32(input_data[2 * w_in - 1]);
+
+      tmp0 = vextq_f32(in0, pad0, 1);
+      tmp1 = vextq_f32(in0, pad0, 2);
+      tmp2 = vextq_f32(in2, pad1, 2);
+      tmp3 = vextq_f32(in2, pad1, 2);
+
+      sum = vaddq_f32(in0, tmp0);
+      sum = vaddq_f32(sum, tmp1);
+      sum = vaddq_f32(sum, in2);
+      sum = vaddq_f32(sum, tmp2);
+      sum = vaddq_f32(sum, tmp3);
+      out0 = vmulq_f32(sum, v_coef);
+
+      for (int i = 0; i < c_mid; ++i) {
+        if (i == 0) {
+          vst1q_lane_f32(output_ptr + i, out0, 0);
+        }
+        if (i == 1) {
+          vst1q_lane_f32(output_ptr + i, out0, 1);
+        }
+        if (i == 2) {
+          vst1q_lane_f32(output_ptr + i, out0, 2);
+        }
+      }
+
+      // bottom_right remain
+      float32x4_t pad2 = vdupq_n_f32(input_data[(h_in - 1) * w_in - 1]);
+      float32x4_t pad3 = vdupq_n_f32(input_data[h_in * w_in - 1]);
+
+      tmp0 = vextq_f32(in4, pad2, 1);
+      tmp1 = vextq_f32(in4, pad2, 2);
+      tmp2 = vextq_f32(in6, pad3, 2);
+      tmp3 = vextq_f32(in6, pad3, 2);
+
+      sum = vaddq_f32(in4, tmp0);
+      sum = vaddq_f32(sum, tmp1);
+      sum = vaddq_f32(sum, in6);
+      sum = vaddq_f32(sum, tmp2);
+      sum = vaddq_f32(sum, tmp3);
+      out0 = vmulq_f32(sum, v_coef);
+
+      for (int i = 0; i < c_mid; ++i) {
+        if (i == 0) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, out0, 0);
+        }
+        if (i == 1) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, out0, 1);
+        }
+        if (i == 2) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, out0, 2);
+        }
+      }
+      // mid
+      for (int j = 0; j < h_out - 2; ++j) {
+        output_ptr = out_data + w_out * (j + 1) + 1;
+        input_tmp = input_data + j * w_in;
+
+        in0 = vld1q_f32(input_tmp);
+        in2 = vld1q_f32(input_tmp + w_in);
+        in4 = vld1q_f32(input_tmp + 2 * w_in);
+        c_mid = w_out - 2;
+        for (; c_mid > 3; c_mid -= 4) {
+          in1 = vld1q_f32(input_tmp + 4);
+          in3 = vld1q_f32(input_tmp + w_in + 4);
+          in5 = vld1q_f32(input_tmp + 2 * w_in + 4);
+
+          tmp0 = vextq_f32(in0, in1, 1);
+          tmp1 = vextq_f32(in0, in1, 2);
+          tmp2 = vextq_f32(in2, in3, 1);
+          tmp3 = vextq_f32(in2, in3, 2);
+          tmp4 = vextq_f32(in4, in5, 1);
+          tmp5 = vextq_f32(in4, in5, 2);
+
+          sum = vaddq_f32(in0, tmp0);
+          sum = vaddq_f32(sum, tmp1);
+          sum = vaddq_f32(sum, in2);
+          sum = vaddq_f32(sum, tmp2);
+          sum = vaddq_f32(sum, tmp3);
+          sum = vaddq_f32(sum, in4);
+          sum = vaddq_f32(sum, tmp4);
+          sum = vaddq_f32(sum, tmp5);
+
+          out0 = vmulq_f32(sum, v_coef);
+          vst1q_f32(output_ptr, out0);
+          output_ptr += 4;
+          input_tmp += 4;
+          in0 = in1;
+          in2 = in3;
+          in4 = in5;
+        }
+        // mid remain
+        float32x4_t pad0 = vdupq_n_f32(input_data[(j + 1) * w_in - 1]);
+        float32x4_t pad1 = vdupq_n_f32(input_data[(j + 2) * w_in - 1]);
+        float32x4_t pad2 = vdupq_n_f32(input_data[(j + 2) * w_in - 1]);
+
+        tmp0 = vextq_f32(in0, pad0, 1);
+        tmp1 = vextq_f32(in0, pad0, 2);
+        tmp2 = vextq_f32(in2, pad1, 1);
+        tmp3 = vextq_f32(in2, pad1, 2);
+        tmp4 = vextq_f32(in4, pad2, 1);
+        tmp5 = vextq_f32(in4, pad2, 2);
+
+        sum = vaddq_f32(in0, tmp0);
+        sum = vaddq_f32(sum, tmp1);
+        sum = vaddq_f32(sum, in2);
+        sum = vaddq_f32(sum, tmp2);
+        sum = vaddq_f32(sum, tmp3);
+        sum = vaddq_f32(sum, in4);
+        sum = vaddq_f32(sum, tmp4);
+        sum = vaddq_f32(sum, tmp5);
+        out0 = vmulq_f32(sum, v_coef);
+
+        for (int i = 0; i < c_mid; ++i) {
+          if (i == 0) {
+            vst1q_lane_f32(output_ptr + i, out0, 0);
+          }
+          if (i == 1) {
+            vst1q_lane_f32(output_ptr + i, out0, 1);
+          }
+          if (i == 2) {
+            vst1q_lane_f32(output_ptr + i, out0, 2);
+          }
+        }
+      }
+      input_data += inputdata_channel_stride;
+      out_data += outputdata_channel_stride;
+    }
+  }
+#endif
+}
+
+void Pool3x3Maxs1p1(const Tensor *input, Tensor *output) {
+#if __ARM_NEON
+  const int batch_size = input->dims()[0];
+
+  const int h_in = input->dims()[2];
+
+  const int w_in = input->dims()[3];
+
+  const int output_channels = output->dims()[1];
+
+  const int h_out = output->dims()[2];
+  const int w_out = output->dims()[3];
+  const int outputdata_channel_stride = h_out * w_out;
+  const int inputdata_channel_stride = h_in * w_in;
+  float *out_data = output->data<float>();
+  const float *input_data = input->data<float>();
+  for (int k = 0; k < batch_size; ++k) {
+    for (int c = 0; c < output_channels; ++c) {
+      // four corner point
+      out_data[0] = std::max(std::max(input_data[0], input_data[1]),
+                             std::max(input_data[w_in], input_data[w_in + 1]));
+      out_data[w_out - 1] = std::max(
+          std::max(input_data[w_in - 2], input_data[w_in - 1]),
+          std::max(input_data[w_in * 2 - 2], input_data[2 * w_in - 1]));
+      out_data[(h_out - 1) * w_out] =
+          std::max(std::max(input_data[(h_in - 2) * w_in],
+                            input_data[(h_in - 2) * w_in + 1]),
+                   std::max(input_data[(h_in - 1) * w_in],
+                            input_data[(h_in - 1) * w_in + 1]));
+      out_data[h_out * w_out - 1] = std::max(
+          std::max(input_data[(h_in - 1) * w_in - 1],
+                   input_data[(h_in - 1) * w_in - 2]),
+          std::max(input_data[h_in * w_in - 1], input_data[h_in * w_in - 2]));
+      // left side & right side
+      for (int i = 1; i < h_in - 1; ++i) {
+        float max1 = std::max(input_data[i * w_in - w_in],
+                              input_data[i * w_in - w_in + 1]);
+        float max2 = std::max(input_data[i * w_in], input_data[i * w_in + 1]);
+        float max3 = std::max(input_data[i * w_in + w_in],
+                              input_data[i * w_in + w_in + 1]);
+        out_data[i * w_out] = std::max(std::max(max1, max2), max3);
+
+        max1 = std::max(input_data[i * w_in - w_in + w_in - 2],
+                        input_data[i * w_in - w_in + 1 + w_in - 2]);
+        max2 = std::max(input_data[i * w_in + w_in - 2],
+                        input_data[i * w_in + 1 + w_in - 2]);
+        max3 = std::max(input_data[i * w_in + w_in + w_in - 2],
+                        input_data[i * w_in + w_in + 1 + w_in - 2]);
+        out_data[i * w_out + w_out - 1] = std::max(std::max(max1, max2), max3);
+      }
+      // top 1 row & bottom 1 row
+      const float *input_tmp = input_data;
+
+      float32x4_t in0, in1, in2, in3, in4, in5, in6, in7, tmp0, tmp1, tmp2,
+          tmp3, tmp4, tmp5, max;
+      in0 = vld1q_f32(input_tmp);
+      in2 = vld1q_f32(input_tmp + w_in);
+      const float *input_tmp_end = input_tmp + (h_in - 2) * w_in;
+      in4 = vld1q_f32(input_tmp_end);
+      in6 = vld1q_f32(input_tmp_end + w_in);
+      int c_mid = w_out - 2;
+      auto output_ptr = out_data + 1;
+      for (; c_mid > 3; c_mid -= 4) {
+        in1 = vld1q_f32(input_tmp + 4);
+        in3 = vld1q_f32(input_tmp + w_in + 4);
+
+        tmp0 = vextq_f32(in0, in1, 1);
+        tmp1 = vextq_f32(in0, in1, 2);
+
+        tmp2 = vextq_f32(in2, in3, 1);
+        tmp3 = vextq_f32(in2, in3, 2);
+
+        max = vmaxq_f32(in0, tmp0);
+        max = vmaxq_f32(max, tmp1);
+        max = vmaxq_f32(max, in2);
+        max = vmaxq_f32(max, tmp2);
+        max = vmaxq_f32(max, tmp3);
+
+        vst1q_f32(output_ptr, max);
+
+        in5 = vld1q_f32(input_tmp_end + 4);
+        in7 = vld1q_f32(input_tmp_end + w_in + 4);
+
+        tmp0 = vextq_f32(in4, in5, 1);
+        tmp1 = vextq_f32(in4, in5, 2);
+        tmp2 = vextq_f32(in6, in7, 1);
+        tmp3 = vextq_f32(in6, in7, 2);
+
+        max = vmaxq_f32(in4, tmp0);
+        max = vmaxq_f32(max, tmp1);
+        max = vmaxq_f32(max, in6);
+        max = vmaxq_f32(max, tmp2);
+        max = vmaxq_f32(max, tmp3);
+
+        vst1q_f32(output_ptr + (h_out - 1) * w_out, max);
+
+        input_tmp += 4;
+        input_tmp_end += 4;
+        output_ptr += 4;
+        in0 = in1;
+        in2 = in3;
+        in4 = in5;
+        in6 = in7;
+      }
+      // top right remain
+      float32x4_t pad0 = vdupq_n_f32(input_data[w_in - 1]);
+      float32x4_t pad1 = vdupq_n_f32(input_data[2 * w_in - 1]);
+
+      tmp0 = vextq_f32(in0, pad0, 1);
+      tmp1 = vextq_f32(in0, pad0, 2);
+      tmp2 = vextq_f32(in2, pad1, 1);
+      tmp3 = vextq_f32(in2, pad1, 2);
+
+      max = vmaxq_f32(in0, tmp0);
+      max = vmaxq_f32(max, tmp1);
+      max = vmaxq_f32(max, in2);
+      max = vmaxq_f32(max, tmp2);
+      max = vmaxq_f32(max, tmp3);
+
+      for (int i = 0; i < c_mid; ++i) {
+        if (i == 0) {
+          vst1q_lane_f32(output_ptr + i, max, 0);
+        }
+        if (i == 1) {
+          vst1q_lane_f32(output_ptr + i, max, 1);
+        }
+        if (i == 2) {
+          vst1q_lane_f32(output_ptr + i, max, 2);
+        }
+      }
+
+      // bottom_right remain
+      float32x4_t pad2 = vdupq_n_f32(input_data[(h_in - 1) * w_in - 1]);
+      float32x4_t pad3 = vdupq_n_f32(input_data[h_in * w_in - 1]);
+
+      tmp0 = vextq_f32(in4, pad2, 1);
+      tmp1 = vextq_f32(in4, pad2, 2);
+      tmp2 = vextq_f32(in6, pad3, 1);
+      tmp3 = vextq_f32(in6, pad3, 2);
+
+      max = vmaxq_f32(in4, tmp0);
+      max = vmaxq_f32(max, tmp1);
+      max = vmaxq_f32(max, in6);
+      max = vmaxq_f32(max, tmp2);
+      max = vmaxq_f32(max, tmp3);
+
+      for (int i = 0; i < c_mid; ++i) {
+        if (i == 0) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, max, 0);
+        }
+        if (i == 1) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, max, 1);
+        }
+        if (i == 2) {
+          vst1q_lane_f32(output_ptr + (h_out - 1) * w_out + i, max, 2);
+        }
+      }
+      // mid
+      for (int j = 0; j < h_out - 2; ++j) {
+        output_ptr = out_data + (j + 1) * w_out + 1;
+        input_tmp = input_data + j * w_in;
+
+        in0 = vld1q_f32(input_tmp);
+        in2 = vld1q_f32(input_tmp + w_in);
+        in4 = vld1q_f32(input_tmp + 2 * w_in);
+        c_mid = w_out - 2;
+        for (; c_mid > 3; c_mid -= 4) {
+          in1 = vld1q_f32(input_tmp + 4);
+          in3 = vld1q_f32(input_tmp + w_in + 4);
+          in5 = vld1q_f32(input_tmp + 2 * w_in + 4);
+
+          tmp0 = vextq_f32(in0, in1, 1);
+          tmp1 = vextq_f32(in0, in1, 2);
+          tmp2 = vextq_f32(in2, in3, 1);
+          tmp3 = vextq_f32(in2, in3, 2);
+          tmp4 = vextq_f32(in4, in5, 1);
+          tmp5 = vextq_f32(in4, in5, 2);
+
+          max = vmaxq_f32(in0, tmp0);
+          max = vmaxq_f32(max, tmp1);
+          max = vmaxq_f32(max, in2);
+          max = vmaxq_f32(max, tmp2);
+          max = vmaxq_f32(max, tmp3);
+          max = vmaxq_f32(max, in4);
+          max = vmaxq_f32(max, tmp4);
+          max = vmaxq_f32(max, tmp5);
+
+          vst1q_f32(output_ptr, max);
+          output_ptr += 4;
+          input_tmp += 4;
+          in0 = in1;
+          in2 = in3;
+          in4 = in5;
+        }
+        // mid remain
+        float32x4_t pad0 = vdupq_n_f32(input_data[(j + 1) * w_in - 1]);
+        float32x4_t pad1 = vdupq_n_f32(input_data[(j + 2) * w_in - 1]);
+        float32x4_t pad2 = vdupq_n_f32(input_data[(j + 3) * w_in - 1]);
+
+        tmp0 = vextq_f32(in0, pad0, 1);
+        tmp1 = vextq_f32(in0, pad0, 2);
+        tmp2 = vextq_f32(in2, pad1, 1);
+        tmp3 = vextq_f32(in2, pad1, 2);
+        tmp4 = vextq_f32(in4, pad2, 1);
+        tmp5 = vextq_f32(in4, pad2, 2);
+
+        max = vmaxq_f32(in0, tmp0);
+        max = vmaxq_f32(max, tmp1);
+        max = vmaxq_f32(max, in2);
+        max = vmaxq_f32(max, tmp2);
+        max = vmaxq_f32(max, tmp3);
+        max = vmaxq_f32(max, in4);
+        max = vmaxq_f32(max, tmp4);
+        max = vmaxq_f32(max, tmp5);
+
+        for (int i = 0; i < c_mid; ++i) {
+          if (i == 0) {
+            vst1q_lane_f32(output_ptr + i, max, 0);
+          }
+          if (i == 1) {
+            vst1q_lane_f32(output_ptr + i, max, 1);
+          }
+          if (i == 2) {
+            vst1q_lane_f32(output_ptr + i, max, 2);
+          }
+        }
+      }
+      input_data += inputdata_channel_stride;
+      out_data += outputdata_channel_stride;
+    }
+  }
+#endif
+}
 
 void Pool3x3Max(vector<int> strides, vector<int> paddings, const Tensor *input,
                 Tensor *output) {

src/operators/math/pool_3x3.h

@@ -15,7 +15,8 @@ limitations under the License. */
 #ifdef POOL_OP
 
 #pragma once
-
+#include <algorithm>
+#include <vector>
 #include "framework/tensor.h"
 #if __ARM_NEON
 #include <arm_neon.h>
@@ -26,7 +27,8 @@ namespace operators {
 namespace math {
 using framework::Tensor;
 using std::vector;
-
+void Pool3x3Avgs1p1(const Tensor *input, Tensor *output);
+void Pool3x3Maxs1p1(const Tensor *input, Tensor *output);
 void Pool3x3Max(vector<int> strides, vector<int> paddings, const Tensor *input,
                 Tensor *output);
 

src/operators/op_param.h

@@ -195,8 +195,7 @@ class OpParam {
 class ConvParam : OpParam {
  public:
   ConvParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-            const framework::AttributeMap &attrs,
-            const framework::Scope &scope) {
+            const AttributeMap &attrs, const Scope &scope) {
     filter_ = FilterFrom<LoDTensor>(inputs, scope);
     input_ = InputFrom<LoDTensor>(inputs, scope);
     output_ = OutputFrom<LoDTensor>(outputs, scope);
@@ -237,12 +236,11 @@ Print &operator<<(Print &printer, const ConvParam &conv_param);
 class ElementwiseAddParam : OpParam {
  public:
   ElementwiseAddParam(const VariableNameMap &inputs,
-                      const VariableNameMap &outputs,
-                      const framework::AttributeMap &attrs,
-                      const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    input_y_ = InputYFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+                      const VariableNameMap &outputs, const AttributeMap &attrs,
+                      const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    input_y_ = InputYFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
     axis_ = GetAttr<int>("axis", attrs);
   }
 
@@ -267,11 +265,10 @@ class ElementwiseAddParam : OpParam {
 class MulParam : OpParam {
  public:
   MulParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-           const framework::AttributeMap &attrs,
-           const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    input_y_ = InputYFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+           const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    input_y_ = InputYFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
     x_num_col_dims_ = GetAttr<int>("x_num_col_dims", attrs);
     y_num_col_dims_ = GetAttr<int>("y_num_col_dims", attrs);
   }
@@ -299,10 +296,9 @@ class MulParam : OpParam {
 class ConcatParam : public OpParam {
  public:
   ConcatParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-              const framework::AttributeMap &attrs,
-              const framework::Scope &scope) {
+              const AttributeMap &attrs, const Scope &scope) {
     inputs_ = InputMultiFrom<LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
     axis_ = GetAttr<int>("axis", attrs);
   }
 
@@ -323,11 +319,10 @@ class ConcatParam : public OpParam {
 class LrnParam : public OpParam {
  public:
   LrnParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-           const framework::AttributeMap &attrs,
-           const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
-    mid_out_ = MidOutFrom<framework::LoDTensor>(outputs, scope);
+           const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
+    mid_out_ = MidOutFrom<LoDTensor>(outputs, scope);
     n_ = GetAttr<int>("n", attrs);
     alpha_ = GetAttr<float>("alpha", attrs);
     beta_ = GetAttr<float>("beta", attrs);
@@ -367,14 +362,13 @@ class LrnParam : public OpParam {
 class BatchNormParam : OpParam {
  public:
   BatchNormParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-                 const framework::AttributeMap &attrs,
-                 const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    output_y_ = OutputYFrom<framework::LoDTensor>(outputs, scope);
-    input_bias_ = InputBiasFrom<framework::LoDTensor>(inputs, scope);
-    input_mean_ = InputMeanFrom<framework::LoDTensor>(inputs, scope);
-    input_scale_ = InputScaleFrom<framework::LoDTensor>(inputs, scope);
-    input_variance_ = InputVarianceFrom<framework::LoDTensor>(inputs, scope);
+                 const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    output_y_ = OutputYFrom<LoDTensor>(outputs, scope);
+    input_bias_ = InputBiasFrom<LoDTensor>(inputs, scope);
+    input_mean_ = InputMeanFrom<LoDTensor>(inputs, scope);
+    input_scale_ = InputScaleFrom<LoDTensor>(inputs, scope);
+    input_variance_ = InputVarianceFrom<LoDTensor>(inputs, scope);
     epsilon_ = GetAttr<float>("epsilon", attrs);
     momentum_ = GetAttr<float>("momentum", attrs);
     is_test_ = GetAttr<bool>("is_test", attrs);
@@ -418,11 +412,10 @@ class BatchNormParam : OpParam {
 class PoolParam : public OpParam {
  public:
   PoolParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-            const framework::AttributeMap &attrs,
-            const framework::Scope &scope) {
-    input_ = InputXFrom<framework::LoDTensor>(inputs, scope);
+            const AttributeMap &attrs, const Scope &scope) {
+    input_ = InputXFrom<LoDTensor>(inputs, scope);
 
-    output_ = OutFrom<framework::LoDTensor>(outputs, scope);
+    output_ = OutFrom<LoDTensor>(outputs, scope);
     pooling_type_ = GetAttr<string>("pooling_type", attrs);
     ksize_ = GetAttr<vector<int>>("ksize", attrs);
     strides_ = GetAttr<vector<int>>("strides", attrs);
@@ -464,13 +457,11 @@ class PoolParam : public OpParam {
 class PriorBoxParam : public OpParam {
  public:
   PriorBoxParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-                const framework::AttributeMap &attrs,
-                const framework::Scope &scope) {
-    input_ = InputFrom<framework::LoDTensor>(inputs, scope);
-    input_image_ = InputImageFrom<framework::LoDTensor>(inputs, scope);
-    output_boxes_ = OutputBoxesFrom<framework::LoDTensor>(outputs, scope);
-    output_variances_ =
-        OutputVariancesFrom<framework::LoDTensor>(outputs, scope);
+                const AttributeMap &attrs, const Scope &scope) {
+    input_ = InputFrom<LoDTensor>(inputs, scope);
+    input_image_ = InputImageFrom<LoDTensor>(inputs, scope);
+    output_boxes_ = OutputBoxesFrom<LoDTensor>(outputs, scope);
+    output_variances_ = OutputVariancesFrom<LoDTensor>(outputs, scope);
     min_sizes_ = GetAttr<vector<float>>("min_sizes", attrs);
     max_sizes_ = GetAttr<vector<float>>("max_sizes", attrs);
     aspect_ratios_ = GetAttr<vector<float>>("aspect_ratios", attrs);
@@ -528,13 +519,11 @@ class PriorBoxParam : public OpParam {
 class BoxCoderParam : public OpParam {
  public:
   BoxCoderParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-                const framework::AttributeMap &attrs,
-                const framework::Scope &scope) {
-    input_priorbox_ = InputPriorBoxFrom<framework::LoDTensor>(inputs, scope);
-    input_priorboxvar_ =
-        InputPriorBoxVarFrom<framework::LoDTensor>(inputs, scope);
-    input_targetbox_ = InputTargetBoxFrom<framework::LoDTensor>(inputs, scope);
-    output_box_ = OutputBoxFrom<framework::LoDTensor>(outputs, scope);
+                const AttributeMap &attrs, const Scope &scope) {
+    input_priorbox_ = InputPriorBoxFrom<LoDTensor>(inputs, scope);
+    input_priorboxvar_ = InputPriorBoxVarFrom<LoDTensor>(inputs, scope);
+    input_targetbox_ = InputTargetBoxFrom<LoDTensor>(inputs, scope);
+    output_box_ = OutputBoxFrom<LoDTensor>(outputs, scope);
     code_type_ = GetAttr<std::string>("code_type", attrs);
   }
   const Tensor *InputPriorBox() const { return input_priorbox_; }
@@ -560,10 +549,9 @@ class BoxCoderParam : public OpParam {
 class SoftmaxParam : public OpParam {
  public:
   SoftmaxParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-               const framework::AttributeMap &attrs,
-               const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+               const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
   }
   const Tensor *InputX() const { return input_x_; }
   Tensor *Out() const { return out_; }
@@ -578,10 +566,9 @@ class SoftmaxParam : public OpParam {
 class SigmoidParam : public OpParam {
  public:
   SigmoidParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-               const framework::AttributeMap &attrs,
-               const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+               const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
   }
   const Tensor *InputX() const { return input_x_; }
   Tensor *Out() const { return out_; }
@@ -643,9 +630,9 @@ class MultiClassNMSParam : public OpParam {
 class FeedParam : public OpParam {
  public:
   FeedParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-            const framework::AttributeMap &attrs, framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+            const AttributeMap &attrs, Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
     auto var = scope.Var("batch_size");
     batch_size = var->GetValue<int>();
   }
@@ -662,10 +649,9 @@ class FeedParam : public OpParam {
 class FetchParam : public OpParam {
  public:
   FetchParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
-             const framework::AttributeMap &attrs,
-             const framework::Scope &scope) {
-    input_x_ = InputXFrom<framework::LoDTensor>(inputs, scope);
-    out_ = OutFrom<framework::LoDTensor>(outputs, scope);
+             const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<LoDTensor>(inputs, scope);
+    out_ = OutFrom<LoDTensor>(outputs, scope);
   }
   const Tensor *InputX() const { return input_x_; }
   Tensor *Out() const { return out_; }
@@ -863,10 +849,10 @@ class FusionConvAddBNReluParam : public OpParam {
     paddings_ = GetAttr<vector<int>>("paddings", attrs);
     dilations_ = GetAttr<vector<int>>("dilations", attrs);
     groups = GetAttr<int>("groups", attrs);
-    input_bias_ = InputBiasFrom<framework::LoDTensor>(inputs, scope);
-    input_mean_ = InputMeanFrom<framework::LoDTensor>(inputs, scope);
-    input_scale_ = InputScaleFrom<framework::LoDTensor>(inputs, scope);
-    input_variance_ = InputVarianceFrom<framework::LoDTensor>(inputs, scope);
+    input_bias_ = InputBiasFrom<LoDTensor>(inputs, scope);
+    input_mean_ = InputMeanFrom<LoDTensor>(inputs, scope);
+    input_scale_ = InputScaleFrom<LoDTensor>(inputs, scope);
+    input_variance_ = InputVarianceFrom<LoDTensor>(inputs, scope);
     epsilon_ = GetAttr<float>("epsilon", attrs);
     momentum_ = GetAttr<float>("momentum", attrs);
     is_test_ = GetAttr<bool>("is_test", attrs);

test/net/test_googlenet.cpp

@@ -17,25 +17,25 @@ limitations under the License. */
 #include "../test_include.h"
 
 int main() {
-  paddle_mobile::PaddleMobile<paddle_mobile::CPU> paddle_mobile;
+  paddle_mobile::Loader<paddle_mobile::CPU> loader;
   bool optimize = true;
   auto time1 = time();
   //  auto program = loader.Load(g_googlenet, optimize);
-  if (paddle_mobile.Load(g_googlenet_combine + "/model",
-                         g_googlenet_combine + "/params", optimize)) {
+  auto program = loader.Load(g_googlenet_combine + "/model",
+                             g_googlenet_combine + "/params", optimize);
   auto time2 = time();
   DLOG << "load cost :" << time_diff(time1, time2) << "ms\n";
+  paddle_mobile::Executor<paddle_mobile::CPU> executor(program, 1, optimize);
   std::vector<float> input;
   std::vector<int64_t> dims{1, 3, 224, 224};
   GetInput<float>(g_test_image_1x3x224x224, &input, dims);
   auto time3 = time();
 
   for (int i = 0; i < 10; ++i) {
-      paddle_mobile.Predict(input, dims);
+    executor.Predict(input, dims);
   }
 
   auto time4 = time();
   DLOG << "predict cost :" << time_diff(time3, time4) << "ms\n";
-  }
   return 0;
 }

tools/build.sh

@@ -32,8 +32,8 @@ build_for_mac() {
 
 build_for_android() {
     #rm -rf "../build"
-    if [ -z "${ANDROID_NDK}" ]; then
-        echo "ANDROID_NDK not found!"
+    if [ -z "${NDK_ROOT}" ]; then
+        echo "NDK_ROOT not found!"
         exit -1
     fi