fix conflict

cmake/external/mkldnn.cmake

@@ -51,7 +51,7 @@ ExternalProject_Add(
     ${EXTERNAL_PROJECT_LOG_ARGS}
     DEPENDS             ${MKLDNN_DEPENDS}
     GIT_REPOSITORY      "https://github.com/01org/mkl-dnn.git"
-    GIT_TAG             "v0.9"
+    GIT_TAG             "v0.10"
     PREFIX              ${MKLDNN_SOURCES_DIR}
     UPDATE_COMMAND      ""
     CMAKE_ARGS          -DCMAKE_INSTALL_PREFIX=${MKLDNN_INSTALL_DIR}

cmake/external/mklml.cmake

@@ -28,7 +28,7 @@ INCLUDE(ExternalProject)
 
 SET(MKLML_PROJECT       "extern_mklml")
 SET(MKLML_VER           "mklml_lnx_2018.0.20170720")
-SET(MKLML_URL           "https://github.com/01org/mkl-dnn/releases/download/v0.9/${MKLML_VER}.tgz")
+SET(MKLML_URL           "https://github.com/01org/mkl-dnn/releases/download/v0.10/${MKLML_VER}.tgz")
 SET(MKLML_SOURCE_DIR    "${THIRD_PARTY_PATH}/mklml")
 SET(MKLML_DOWNLOAD_DIR  "${MKLML_SOURCE_DIR}/src/${MKLML_PROJECT}")
 SET(MKLML_DST_DIR       "mklml")
@@ -54,7 +54,8 @@ ExternalProject_Add(
     ${EXTERNAL_PROJECT_LOG_ARGS}
     PREFIX                ${MKLML_SOURCE_DIR}
     DOWNLOAD_DIR          ${MKLML_DOWNLOAD_DIR}
-    DOWNLOAD_COMMAND      wget --no-check-certificate -qO- ${MKLML_URL} | tar xz -C ${MKLML_DOWNLOAD_DIR}
+    DOWNLOAD_COMMAND      wget --no-check-certificate ${MKLML_URL} -c -q -O ${MKLML_VER}.tgz 
+                          && tar zxf ${MKLML_VER}.tgz
     DOWNLOAD_NO_PROGRESS  1
     UPDATE_COMMAND        ""
     CMAKE_ARGS            -DCMAKE_INSTALL_PREFIX=${MKLML_INSTALL_ROOT}

doc/getstarted/build_and_install/index_cn.rst

@@ -6,14 +6,12 @@
 安装流程
 ++++++++
 
-PaddlePaddle提供数个预编译的二进制来进行安装，包括Docker镜像，ubuntu的deb安装包等。我们推荐使用Docker镜像来部署环境，同时欢迎贡献更多的安装包。
+PaddlePaddle提供Docker镜像来部署环境。
 
 .. toctree::
    :maxdepth: 1
    
    docker_install_cn.rst 
-   ubuntu_install_cn.rst
-
 
 
 编译流程

doc/getstarted/build_and_install/index_en.rst

@@ -8,14 +8,13 @@ Install PaddlePaddle
     :maxdepth: 1
 
     docker_install_en.rst
-    ubuntu_install_en.rst
 
 Build from Source
 -----------------
 
 ..  warning::
 
-    Please use :code:`deb` package or :code:`docker` image to install paddle. The building guide is used for hacking or contributing PaddlePaddle source code.
+    Please use :code:`docker` image to install paddle. The building guide is used for hacking or contributing PaddlePaddle source code.
 
 ..  toctree::
     :maxdepth: 1

doc/getstarted/build_and_install/ubuntu_install_cn.rst

-Ubuntu部署PaddlePaddle
-===================================
-
-PaddlePaddle提供了ubuntu 14.04 deb安装包。
-
-安装
-------
-
-安装包的下载地址是\: https://github.com/PaddlePaddle/Paddle/releases
-
-它包含四个版本\:
-
-* cpu版本: 支持主流x86处理器平台, 使用了avx指令集。
-
-* cpu-noavx版本：支持主流x86处理器平台，没有使用avx指令集。
-
-* gpu版本：支持主流x86处理器平台，支持nvidia cuda平台，使用了avx指令集。
-
-* gpu-noavx版本：支持主流x86处理器平台，支持nvidia cuda平台，没有使用avx指令集。
-
-下载完相关安装包后，执行:
-
-..  code-block:: shell
-
-    sudo apt-get install gdebi
-    gdebi paddle-*-cpu.deb
-
-或者:
-
-..  code-block:: shell
-
-    dpkg -i paddle-*-cpu.deb
-    apt-get install -f
-
-
-在 :code:`dpkg -i` 的时候如果报一些依赖未找到的错误是正常的，
-在 :code:`apt-get install -f` 里会继续安装 PaddlePaddle。
-
-安装完成后，可以使用命令 :code:`paddle version` 查看安装后的paddle 版本:
-
-..  code-block:: shell
-
-    PaddlePaddle 0.8.0b1, compiled with
-        with_avx: ON
-        with_gpu: OFF
-        with_double: OFF
-        with_python: ON
-        with_rdma: OFF
-        with_timer: OFF
-        with_predict_sdk:
-
-
-可能遇到的问题
---------------
-
-libcudart.so/libcudnn.so找不到
-++++++++++++++++++++++++++++++
-
-安装完成后，运行 :code:`paddle train` 报错\:
-
-..  code-block:: shell
-
-      0831 12:36:04.151525  1085 hl_dso_loader.cc:70] Check failed: nullptr != *dso_handle For Gpu version of PaddlePaddle, it couldn't find CUDA library: libcudart.so Please make sure you already specify its path.Note: for training data on Cpu using Gpu version of PaddlePaddle,you must specify libcudart.so via LD_LIBRARY_PATH.
-
-原因是未设置cuda运行时环境变量。 如果使用GPU版本的PaddlePaddle，请安装CUDA 7.5 和CUDNN 5到本地环境中，并设置：
-
-..  code-block:: shell
-
-    export LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib:$LD_LIBRARY_PATH
-    export PATH=/usr/local/cuda/bin:$PATH
-

doc/getstarted/build_and_install/ubuntu_install_en.rst

-Debian Package installation guide
-=================================
-
-PaddlePaddle supports :code:`deb` pacakge. The installation of this :code:`deb` package is tested in ubuntu 14.04, but it should be support other debian based linux, too.
-
-There are four versions of debian package, :code:`cpu`, :code:`gpu`, :code:`cpu-noavx`, :code:`gpu-noavx`. And :code:`noavx` version is used to support CPU which does not contain :code:`AVX` instructions. The download url of :code:`deb` package is \: https://github.com/baidu/Paddle/releases/
-
-
-After downloading PaddlePaddle deb packages, you can use :code:`gdebi` install.
-
-..	code-block:: bash
-
-	gdebi paddle-*.deb
-
-If :code:`gdebi` is not installed, you can use :code:`sudo apt-get install gdebi` to install it.
-
-Or you can use following commands to install PaddlePaddle.
-
-..	code-block:: bash
-
-	dpkg -i paddle-*.deb
-	apt-get install -f
-
-And if you use GPU version deb package, you need to install CUDA toolkit and cuDNN, and set related environment variables(such as LD_LIBRARY_PATH) first. It is normal when `dpkg -i` get errors. `apt-get install -f` will continue install paddle, and install dependences. 
-

paddle/function/CMakeLists.txt

@@ -21,6 +21,8 @@ if(USE_NNPACK)
   endif()
 endif()
 
+list(APPEND cpp_files neon/NeonDepthwiseConv.cpp)
+
 add_library(paddle_function STATIC ${cpp_files} ${cu_objs})
 add_dependencies(paddle_function ${external_project_dependencies})
 add_dependencies(paddle_function paddle_proto)
@@ -42,11 +44,11 @@ if(WITH_GPU)
     add_simple_unittest(RowConvOpTest)
     add_simple_unittest(BlockExpandOpTest)
     add_simple_unittest(CropOpTest)
-    add_simple_unittest(DepthwiseConvOpTest)
 endif()
 
 add_simple_unittest(Im2ColTest)
 add_simple_unittest(GemmConvOpTest)
+add_simple_unittest(DepthwiseConvOpTest)
 endif()
 
 add_style_check_target(paddle_function ${h_files})

paddle/function/DepthwiseConvOpTest.cpp

@@ -34,4 +34,13 @@ TEST(DepthwiseConv, BackwardFilter) {
 }
 #endif
 
+#if defined(__ARM_NEON__) || defined(__ARM_NEON)
+
+TEST(DepthwiseConv, Forward) {
+  DepthwiseConvolution<DEVICE_TYPE_CPU, DEVICE_TYPE_CPU>(
+      "GemmConv-CPU", "NeonDepthwiseConv-CPU", forward);
+}
+
+#endif
+
 }  // namespace paddle

paddle/function/Im2Col.h

@@ -16,6 +16,7 @@ limitations under the License. */
 
 #include "TensorShape.h"
 #include "TensorType.h"
+#include "neon/neon_util.h"
 
 namespace paddle {
 
@@ -93,4 +94,95 @@ public:
                   int paddingWidth);
 };
 
+template <class T>
+struct Padding {
+  static void run(const T* src,
+                  T* dest,
+                  int channels,
+                  int inputHeight,
+                  int inputWidth,
+                  int paddingHeight,
+                  int paddingWidth) {
+    const int destWidth = inputWidth + 2 * paddingWidth;
+    for (int c = 0; c < channels; c++) {
+      if (paddingHeight > 0) {
+        memset(dest, 0, destWidth * paddingHeight * sizeof(T));
+        dest += destWidth * paddingHeight;
+      }
+
+      for (int i = 0; i < inputHeight; i++) {
+        // padding head
+        for (int j = 0; j < paddingWidth; j++) {
+          *dest++ = T(0);
+        }
+
+        memcpy(dest, src, inputWidth * sizeof(T));
+        dest += inputWidth;
+        src += inputWidth;
+
+        // padding tail
+        for (int j = 0; j < paddingWidth; j++) {
+          *dest++ = T(0);
+        }
+      }
+
+      if (paddingHeight > 0) {
+        memset(dest, 0, destWidth * paddingHeight * sizeof(T));
+        dest += destWidth * paddingHeight;
+      }
+    }
+  }
+};
+
+#if defined(__ARM_NEON__) || defined(__ARM_NEON)
+template <>
+struct Padding<float> {
+  static void run(const float* src,
+                  float* dest,
+                  int channels,
+                  int inputHeight,
+                  int inputWidth,
+                  int paddingHeight,
+                  int paddingWidth) {
+    const int destWidth = inputWidth + 2 * paddingWidth;
+    for (int c = 0; c < channels; c++) {
+      if (paddingHeight > 0) {
+        memset(dest, 0, destWidth * paddingHeight * sizeof(float));
+        dest += destWidth * paddingHeight;
+      }
+
+      for (int i = 0; i < inputHeight; i++) {
+        // padding head
+        for (int j = 0; j < paddingWidth; j++) {
+          *dest++ = float(0);
+        }
+
+        int step = inputWidth >> 2;
+        int remain = inputWidth & 3;
+        for (int s = 0; s < step; s++) {
+          float32x4_t s0 = vld1q_f32(src);
+          vst1q_f32(dest, s0);
+          src += 4;
+          dest += 4;
+        }
+        for (int r = 0; r < remain; r++) {
+          *dest++ = *src++;
+        }
+
+        // padding tail
+        for (int j = 0; j < paddingWidth; j++) {
+          *dest++ = float(0);
+        }
+      }
+
+      if (paddingHeight > 0) {
+        memset(dest, 0, destWidth * paddingHeight * sizeof(float));
+        dest += destWidth * paddingHeight;
+      }
+    }
+  }
+};
+
+#endif
+
 }  // namespace paddle

paddle/function/neon/NeonDepthwiseConv.cpp

+/* 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. */
+
+#include "neon_util.h"
+#include "paddle/function/ConvOp.h"
+#include "paddle/function/Im2Col.h"
+
+namespace paddle {
+
+namespace neon {
+
+#if defined(__ARM_NEON__) || defined(__ARM_NEON)
+
+template <int filterSize, int stride>
+struct DepthwiseConvKernel {};
+
+inline float32_t conv3x3(float32x4_t r0,
+                         float32x4_t r1,
+                         float32x4_t r2,
+                         float32x4_t k0,
+                         float32x4_t k1,
+                         float32x4_t k2) {
+  float32x4_t tmp;
+  tmp = vmulq_f32(r0, k0);
+  tmp = vmlaq_f32(tmp, r1, k1);
+  tmp = vmlaq_f32(tmp, r2, k2);
+  return vaddvq_f32(tmp);
+}
+
+inline float32_t conv4x4(float32x4_t r0,
+                         float32x4_t r1,
+                         float32x4_t r2,
+                         float32x4_t r3,
+                         float32x4_t k0,
+                         float32x4_t k1,
+                         float32x4_t k2,
+                         float32x4_t k3) {
+  float32x4_t tmp;
+  tmp = vmulq_f32(r0, k0);
+  tmp = vmlaq_f32(tmp, r1, k1);
+  tmp = vmlaq_f32(tmp, r2, k2);
+  tmp = vmlaq_f32(tmp, r3, k3);
+  return vaddvq_f32(tmp);
+}
+
+/**
+ * Each step calculates four elements of the output.
+ * First step:
+ *   R0[0, 1, 2, 3...] * K[0][0]
+ *   R0[1, 2, 3, 4...] * K[0][1]
+ *   R0[2, 3, 4, 5...] * K[0][2]
+ *   R1[0, 1, 2, 3...] * K[1][0]
+ *   R1[1, 2, 3, 4...] * K[1][1]
+ *   R1[2, 3, 4, 5...] * K[1][2]
+ *   R2[0, 1, 2, 3...] * K[2][0]
+ *   R2[1, 2, 3, 4...] * K[2][1]
+ * + R2[2, 3, 4, 5...] * K[2][2]
+ * ------------------------------
+ *     Output[0, 1, 2, 3]
+ */
+template <>
+struct DepthwiseConvKernel<3, 1> {
+  static void run(const float* inputData,
+                  const float* filterData,
+                  int inputHeight,
+                  int inputWidth,
+                  int outputChannels,
+                  int outputHeight,
+                  int outputWidth,
+                  int filterMultiplier,
+                  float* outputData) {
+    const int steps = outputWidth >> 2;
+    const int remain = outputWidth & 3;
+    for (int c = 0; c < outputChannels; c++, filterData += 9) {
+      // Load the filters
+      float32x4_t k[3];
+      k[0] = vld1q_f32(filterData);
+      k[1] = vld1q_f32(filterData + 3);
+      k[2] = vld1q_f32(filterData + 6);
+      k[0] = vsetq_lane_f32(0.f, k[0], 3);
+      k[1] = vsetq_lane_f32(0.f, k[1], 3);
+      k[2] = vsetq_lane_f32(0.f, k[2], 3);
+
+      const float* r0 =
+          inputData + (c / filterMultiplier) * (inputHeight * inputWidth);
+      const float* r1 = r0 + inputWidth;
+      const float* r2 = r0 + inputWidth * 2;
+      float32x4_t input[3][3];
+      for (int h = 0; h < outputHeight; h++) {
+        for (int s = 0; s < steps; s++) {
+          // Load the inputs
+          float32x4_t tmp;
+          input[0][0] = vld1q_f32(r0);
+          tmp = vld1q_f32(r0 + 4);
+          input[0][1] = vextq_f32(input[0][0], tmp, 1);
+          input[0][2] = vextq_f32(input[0][0], tmp, 2);
+          input[1][0] = vld1q_f32(r1);
+          tmp = vld1q_f32(r1 + 4);
+          input[1][1] = vextq_f32(input[1][0], tmp, 1);
+          input[1][2] = vextq_f32(input[1][0], tmp, 2);
+          input[2][0] = vld1q_f32(r2);
+          tmp = vld1q_f32(r2 + 4);
+          input[2][1] = vextq_f32(input[2][0], tmp, 1);
+          input[2][2] = vextq_f32(input[2][0], tmp, 2);
+
+          float32x4_t tmp1 = vdupq_n_f32(0.f);
+          float32x4_t tmp2 = vdupq_n_f32(0.f);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][0], k[0], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][1], k[0], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][2], k[0], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][0], k[1], 0);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][1], k[1], 1);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][2], k[1], 2);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][0], k[2], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][1], k[2], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][2], k[2], 2);
+          tmp1 = vaddq_f32(tmp1, tmp2);
+
+          vst1q_f32(outputData, tmp1);
+          r0 += 4;
+          r1 += 4;
+          r2 += 4;
+          outputData += 4;
+        }
+
+        for (int r = 0; r < remain; r++) {
+          float32x4_t i0 = vld1q_f32(r0);
+          float32x4_t i1 = vld1q_f32(r1);
+          float32x4_t i2 = vld1q_f32(r2);
+          *outputData = conv3x3(i0, i1, i2, k[0], k[1], k[2]);
+          r0++;
+          r1++;
+          r2++;
+          outputData++;
+        }
+
+        r0 += 2;
+        r1 += 2;
+        r2 += 2;
+      }
+    }
+  }
+};
+
+/**
+ * Each step calculates four elements of the output.
+ * First step:
+ *   R0[0, 2, 4, 6...] * K[0][0]
+ *   R0[1, 3, 5, 7...] * K[0][1]
+ *   R0[2, 4, 6, 8...] * K[0][2]
+ *   R1[0, 2, 4, 6...] * K[1][0]
+ *   R1[1, 3, 5, 7...] * K[1][1]
+ *   R1[2, 4, 6, 8...] * K[1][2]
+ *   R2[0, 2, 4, 6...] * K[2][0]
+ *   R2[1, 3, 5, 7...] * K[2][1]
+ *   R2[2, 4, 6, 8...] * K[2][2]
+ * ------------------------------
+ *     Output[0, 1, 2, 3]
+ */
+template <>
+struct DepthwiseConvKernel<3, 2> {
+  static void run(const float* inputData,
+                  const float* filterData,
+                  int inputHeight,
+                  int inputWidth,
+                  int outputChannels,
+                  int outputHeight,
+                  int outputWidth,
+                  int filterMultiplier,
+                  float* outputData) {
+    const int steps = outputWidth >> 2;
+    const int remain = outputWidth & 3;
+    for (int c = 0; c < outputChannels; c++, filterData += 9) {
+      // Load the filters
+      float32x4_t k[3];
+      k[0] = vld1q_f32(filterData);
+      k[1] = vld1q_f32(filterData + 3);
+      k[2] = vld1q_f32(filterData + 6);
+      k[0] = vsetq_lane_f32(0.f, k[0], 3);
+      k[1] = vsetq_lane_f32(0.f, k[1], 3);
+      k[2] = vsetq_lane_f32(0.f, k[2], 3);
+
+      const float* start =
+          inputData + (c / filterMultiplier) * (inputHeight * inputWidth);
+      float32x4_t input[3][3];
+      for (int h = 0; h < outputHeight; h++) {
+        const float* r0 = start + 2 * h * inputWidth;
+        const float* r1 = start + (2 * h + 1) * inputWidth;
+        const float* r2 = start + (2 * h + 2) * inputWidth;
+        for (int s = 0; s < steps; s++) {
+          // Load the inputs
+          float32x4_t data1;
+          float32x4x2_t data2;
+
+          data2 = vld2q_f32(r0);
+          input[0][0] = data2.val[0];
+          input[0][1] = data2.val[1];
+          data1 = vld1q_f32(r0 + 8);
+          input[0][2] = vextq_f32(data2.val[0], data1, 1);
+
+          data2 = vld2q_f32(r1);
+          input[1][0] = data2.val[0];
+          input[1][1] = data2.val[1];
+          data1 = vld1q_f32(r1 + 8);
+          input[1][2] = vextq_f32(data2.val[0], data1, 1);
+
+          data2 = vld2q_f32(r2);
+          input[2][0] = data2.val[0];
+          input[2][1] = data2.val[1];
+          data1 = vld1q_f32(r2 + 8);
+          input[2][2] = vextq_f32(data2.val[0], data1, 1);
+
+          float32x4_t tmp1 = vdupq_n_f32(0.f);
+          float32x4_t tmp2 = vdupq_n_f32(0.f);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][0], k[0], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][1], k[0], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][2], k[0], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][0], k[1], 0);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][1], k[1], 1);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][2], k[1], 2);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][0], k[2], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][1], k[2], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][2], k[2], 2);
+          tmp1 = vaddq_f32(tmp1, tmp2);
+
+          vst1q_f32(outputData, tmp1);
+          r0 += 8;
+          r1 += 8;
+          r2 += 8;
+          outputData += 4;
+        }
+
+        for (int r = 0; r < remain; r++) {
+          float32x4_t i0 = vld1q_f32(r0);
+          float32x4_t i1 = vld1q_f32(r1);
+          float32x4_t i2 = vld1q_f32(r2);
+          *outputData = conv3x3(i0, i1, i2, k[0], k[1], k[2]);
+          r0 += 2;
+          r1 += 2;
+          r2 += 2;
+          outputData++;
+        }
+      }
+    }
+  }
+};
+
+/**
+ * Each step calculates four elements of the output.
+ */
+template <>
+struct DepthwiseConvKernel<4, 1> {
+  static void run(const float* inputData,
+                  const float* filterData,
+                  int inputHeight,
+                  int inputWidth,
+                  int outputChannels,
+                  int outputHeight,
+                  int outputWidth,
+                  int filterMultiplier,
+                  float* outputData) {
+    const int steps = outputWidth >> 2;
+    const int remain = outputWidth & 3;
+    for (int c = 0; c < outputChannels; c++, filterData += 16) {
+      // Load the filters
+      float32x4_t k[4];
+      k[0] = vld1q_f32(filterData);
+      k[1] = vld1q_f32(filterData + 4);
+      k[2] = vld1q_f32(filterData + 8);
+      k[3] = vld1q_f32(filterData + 12);
+
+      const float* r0 =
+          inputData + (c / filterMultiplier) * (inputHeight * inputWidth);
+      const float* r1 = r0 + inputWidth;
+      const float* r2 = r0 + inputWidth * 2;
+      const float* r3 = r0 + inputWidth * 3;
+      float32x4_t input[4][4];
+      for (int h = 0; h < outputHeight; h++) {
+        for (int s = 0; s < steps; s++) {
+          // Load the inputs
+          float32x4_t tmp;
+          input[0][0] = vld1q_f32(r0);
+          tmp = vld1q_f32(r0 + 4);
+          input[0][1] = vextq_f32(input[0][0], tmp, 1);
+          input[0][2] = vextq_f32(input[0][0], tmp, 2);
+          input[0][3] = vextq_f32(input[0][0], tmp, 3);
+
+          input[1][0] = vld1q_f32(r1);
+          tmp = vld1q_f32(r1 + 4);
+          input[1][1] = vextq_f32(input[1][0], tmp, 1);
+          input[1][2] = vextq_f32(input[1][0], tmp, 2);
+          input[1][3] = vextq_f32(input[1][0], tmp, 3);
+
+          input[2][0] = vld1q_f32(r2);
+          tmp = vld1q_f32(r2 + 4);
+          input[2][1] = vextq_f32(input[2][0], tmp, 1);
+          input[2][2] = vextq_f32(input[2][0], tmp, 2);
+          input[2][3] = vextq_f32(input[2][0], tmp, 3);
+
+          input[3][0] = vld1q_f32(r3);
+          tmp = vld1q_f32(r3 + 4);
+          input[3][1] = vextq_f32(input[3][0], tmp, 1);
+          input[3][2] = vextq_f32(input[3][0], tmp, 2);
+          input[3][3] = vextq_f32(input[3][0], tmp, 3);
+
+          float32x4_t tmp1 = vdupq_n_f32(0.f);
+          float32x4_t tmp2 = vdupq_n_f32(0.f);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][0], k[0], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][1], k[0], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][2], k[0], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][3], k[0], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][0], k[1], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][1], k[1], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][2], k[1], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][3], k[1], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][0], k[2], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][1], k[2], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][2], k[2], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][3], k[2], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[3][0], k[3], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[3][1], k[3], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[3][2], k[3], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[3][3], k[3], 3);
+          tmp1 = vaddq_f32(tmp1, tmp2);
+
+          vst1q_f32(outputData, tmp1);
+          r0 += 4;
+          r1 += 4;
+          r2 += 4;
+          r3 += 4;
+          outputData += 4;
+        }
+
+        for (int r = 0; r < remain; r++) {
+          float32x4_t i0 = vld1q_f32(r0);
+          float32x4_t i1 = vld1q_f32(r1);
+          float32x4_t i2 = vld1q_f32(r2);
+          float32x4_t i3 = vld1q_f32(r3);
+          *outputData = conv4x4(i0, i1, i2, i3, k[0], k[1], k[2], k[3]);
+          r0++;
+          r1++;
+          r2++;
+          r3++;
+          outputData++;
+        }
+
+        r0 += 3;
+        r1 += 3;
+        r2 += 3;
+        r3 += 3;
+      }
+    }
+  }
+};
+
+/**
+ * Each step calculates four elements of the output.
+ */
+template <>
+struct DepthwiseConvKernel<4, 2> {
+  static void run(const float* inputData,
+                  const float* filterData,
+                  int inputHeight,
+                  int inputWidth,
+                  int outputChannels,
+                  int outputHeight,
+                  int outputWidth,
+                  int filterMultiplier,
+                  float* outputData) {
+    const int steps = outputWidth >> 2;
+    const int remain = outputWidth & 3;
+    for (int c = 0; c < outputChannels; c++, filterData += 16) {
+      // Load the filters
+      float32x4_t k[4];
+      k[0] = vld1q_f32(filterData);
+      k[1] = vld1q_f32(filterData + 4);
+      k[2] = vld1q_f32(filterData + 8);
+      k[3] = vld1q_f32(filterData + 12);
+
+      const float* start =
+          inputData + (c / filterMultiplier) * (inputHeight * inputWidth);
+      float32x4_t input[4][4];
+      for (int h = 0; h < outputHeight; h++) {
+        const float* r0 = start + 2 * h * inputWidth;
+        const float* r1 = start + (2 * h + 1) * inputWidth;
+        const float* r2 = start + (2 * h + 2) * inputWidth;
+        const float* r3 = start + (2 * h + 3) * inputWidth;
+        for (int s = 0; s < steps; s++) {
+          // Load the inputs
+          float32x4x2_t data1;
+          float32x4x2_t data2;
+
+          data1 = vld2q_f32(r0);
+          data2 = vld2q_f32(r0 + 8);
+          input[0][0] = data1.val[0];
+          input[0][1] = data1.val[1];
+          input[0][2] = vextq_f32(data1.val[0], data2.val[0], 1);
+          input[0][3] = vextq_f32(data1.val[1], data2.val[1], 1);
+
+          data1 = vld2q_f32(r1);
+          data2 = vld2q_f32(r1 + 8);
+          input[1][0] = data1.val[0];
+          input[1][1] = data1.val[1];
+          input[1][2] = vextq_f32(data1.val[0], data2.val[0], 1);
+          input[1][3] = vextq_f32(data1.val[1], data2.val[1], 1);
+
+          data1 = vld2q_f32(r2);
+          data2 = vld2q_f32(r2 + 8);
+          input[2][0] = data1.val[0];
+          input[2][1] = data1.val[1];
+          input[2][2] = vextq_f32(data1.val[0], data2.val[0], 1);
+          input[2][3] = vextq_f32(data1.val[1], data2.val[1], 1);
+
+          data1 = vld2q_f32(r3);
+          data2 = vld2q_f32(r3 + 8);
+          input[3][0] = data1.val[0];
+          input[3][1] = data1.val[1];
+          input[3][2] = vextq_f32(data1.val[0], data2.val[0], 1);
+          input[3][3] = vextq_f32(data1.val[1], data2.val[1], 1);
+
+          float32x4_t tmp1 = vdupq_n_f32(0.f);
+          float32x4_t tmp2 = vdupq_n_f32(0.f);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][0], k[0], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][1], k[0], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[0][2], k[0], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[0][3], k[0], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][0], k[1], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][1], k[1], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[1][2], k[1], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[1][3], k[1], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][0], k[2], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][1], k[2], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[2][2], k[2], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[2][3], k[2], 3);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[3][0], k[3], 0);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[3][1], k[3], 1);
+          tmp1 = vmlaq_laneq_f32(tmp1, input[3][2], k[3], 2);
+          tmp2 = vmlaq_laneq_f32(tmp2, input[3][3], k[3], 3);
+          tmp1 = vaddq_f32(tmp1, tmp2);
+
+          vst1q_f32(outputData, tmp1);
+          r0 += 8;
+          r1 += 8;
+          r2 += 8;
+          r3 += 8;
+          outputData += 4;
+        }
+
+        for (int r = 0; r < remain; r++) {
+          float32x4_t i0 = vld1q_f32(r0);
+          float32x4_t i1 = vld1q_f32(r1);
+          float32x4_t i2 = vld1q_f32(r2);
+          float32x4_t i3 = vld1q_f32(r3);
+          *outputData = conv4x4(i0, i1, i2, i3, k[0], k[1], k[2], k[3]);
+          r0 += 2;
+          r1 += 2;
+          r2 += 2;
+          r3 += 2;
+          outputData++;
+        }
+      }
+    }
+  }
+};
+
+template <DeviceType Device>
+class NeonDepthwiseConvFunction : public ConvFunctionBase {
+public:
+  void init(const FuncConfig& config) override {
+    ConvFunctionBase::init(config);
+  }
+
+  void check(const BufferArgs& inputs, const BufferArgs& outputs) override {
+    const TensorShape& input = inputs[0].shape();
+    const TensorShape& filter = inputs[1].shape();
+    const TensorShape& output = outputs[0].shape();
+    checkShape(input, filter, output);
+  }
+
+  void calc(const BufferArgs& inputs, const BufferArgs& outputs) override {
+    CHECK_EQ(numInputs_, inputs.size());
+    CHECK_EQ(numOutputs_, outputs.size());
+    check(inputs, outputs);
+
+    const TensorShape& input = inputs[0].shape();
+    const TensorShape& filter = inputs[1].shape();
+    const TensorShape& output = outputs[0].shape();
+
+    size_t batchSize = input[0];
+    size_t inputChannels = input[1];
+    size_t inputHeight = input[2];
+    size_t inputWidth = input[3];
+    size_t filterHeight = getFilterHeight(filter);
+    size_t filterWidth = getFilterWidth(filter);
+    size_t outputChannels = output[1];
+    size_t outputHeight = output[2];
+    size_t outputWidth = output[3];
+    size_t filterMultiplier = outputChannels / groups_;
+    CHECK_EQ(inputChannels, groups_);
+
+    // only support strideH() == strideW() and filterHeight == filterWidth.
+    CHECK_EQ(strideH(), strideW());
+    CHECK_EQ(filterHeight, filterWidth);
+
+    float* inputData = inputs[0].data<float>();
+    float* filterData = inputs[1].data<float>();
+    float* outputData = outputs[0].data<float>();
+
+    // padding the input
+    float* inputPadding = inputData;
+    if (paddingH() > 0 || paddingW() > 0) {
+      int newSize = batchSize * inputChannels * (inputHeight + 2 * paddingH()) *
+                    (inputWidth + 2 * paddingW());
+      resizeBuffer<Device>(newSize);
+      inputPadding = reinterpret_cast<float*>(memory_->getBuf());
+      Padding<float>::run(inputData,
+                          inputPadding,
+                          batchSize * inputChannels,
+                          inputHeight,
+                          inputWidth,
+                          paddingH(),
+                          paddingW());
+
+      // height and width of padding data
+      inputHeight += 2 * paddingH();
+      inputWidth += 2 * paddingW();
+    }
+
+    std::function<void(
+        const float*, const float*, int, int, int, int, int, int, float*)>
+        DepthWiseConv;
+
+    if (filterWidth == 3 && strideW() == 1) {
+      DepthWiseConv = DepthwiseConvKernel<3, 1>::run;
+    } else if (filterWidth == 3 && strideW() == 2) {
+      DepthWiseConv = DepthwiseConvKernel<3, 2>::run;
+    } else if (filterWidth == 4 && strideW() == 1) {
+      DepthWiseConv = DepthwiseConvKernel<4, 1>::run;
+    } else if (filterWidth == 4 && strideW() == 2) {
+      DepthWiseConv = DepthwiseConvKernel<4, 2>::run;
+    } else {
+      LOG(FATAL) << "Not supported";
+    }
+
+    for (size_t i = 0; i < batchSize; i++) {
+      DepthWiseConv(inputPadding,
+                    filterData,
+                    inputHeight,
+                    inputWidth,
+                    outputChannels,
+                    outputHeight,
+                    outputWidth,
+                    filterMultiplier,
+                    outputData);
+      inputPadding += inputChannels * inputHeight * inputWidth;
+      outputData += outputChannels * outputHeight * outputWidth;
+    }
+  }
+};
+
+REGISTER_TYPED_FUNC(NeonDepthwiseConv, CPU, NeonDepthwiseConvFunction);
+
+#endif
+
+}  // namespace neon
+}  // namespace paddle

paddle/function/neon/neon_util.h

+/* 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. */
+
+#pragma once
+
+#if defined(__ARM_NEON__) || defined(__ARM_NEON)
+
+#include <arm_neon.h>
+
+namespace paddle {
+
+namespace neon {
+
+inline float32x4_t vld1q_f32_aligned(const float* p) {
+  return vld1q_f32(
+      (const float*)__builtin_assume_aligned(p, sizeof(float32x4_t)));
+}
+
+#ifndef __aarch64__
+inline float32_t vaddvq_f32(float32x4_t a) {
+  float32x2_t v = vadd_f32(vget_high_f32(a), vget_low_f32(a));
+  return vget_lane_f32(vpadd_f32(v, v), 0);
+}
+
+inline float32x4_t vmlaq_laneq_f32(float32x4_t a,
+                                   float32x4_t b,
+                                   float32x4_t v,
+                                   const int lane) {
+  return vmlaq_n_f32(a, b, vgetq_lane_f32(v, lane));
+}
+#endif
+
+}  // namespace neon
+}  // namespace paddle
+
+#endif

paddle/gserver/layers/CostLayer.h

@@ -318,7 +318,9 @@ public:
 
   void forwardImp(Matrix& output, Argument& label, Matrix& cost) override;
 
-  void backwardImp(Matrix& outputValue, Argument& label, Matrix& outputGrad) {}
+  void backwardImp(Matrix& outputValue,
+                   Argument& label,
+                   Matrix& outputGrad) override {}
 };
 
 /**

paddle/gserver/layers/CrossEntropyOverBeam.cpp

+/* 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. */
+
+#include "CrossEntropyOverBeam.h"
+
+namespace paddle {
+
+void CostForOneSequence::calValidExpandStep() {
+  validExpansionCount_ = 0;
+  goldAsExtraPath_ = true;
+
+  for (size_t i = 0; i < beams_->expansionCount; ++i) {
+    real gold = static_cast<real>(beams_->gold[i]);
+    if (i) {
+      real* start = beams_->candidateIds[i - 1]->getData();
+      goldRowIds_[i] = std::count_if(
+          start,
+          start + goldRowIds_[i - 1] * beamSize_ + goldColIds_[i - 1],
+          [](const real& val) { return val != -1.; });
+    } else {
+      goldRowIds_[i] = 0;
+    }
+
+    real* start =
+        beams_->candidateIds[i]->getData() + goldRowIds_[i] * beamSize_;
+    real* findEnd = std::find(start, start + beamSize_, gold);
+    validExpansionCount_++;
+
+    if (start + beamSize_ == findEnd) return;
+    goldColIds_[i] = findEnd - start;
+  }
+  if (goldColIds_[beams_->expansionCount - 1] != -1) goldAsExtraPath_ = false;
+}
+
+size_t CostForOneSequence::initLastExpansion() {
+  int beamId = validExpansionCount_ - 1;
+  const MatrixPtr candidates = beams_->candidateIds[beamId];
+  size_t height = candidates->getHeight();
+
+  /* initialization the last expansion. */
+  size_t pathCount = std::count_if(candidates->getData(),
+                                   candidates->getData() + height * beamSize_,
+                                   [](const real& val) { return val != -1; });
+  /*
+   * if the gold sequence falls off the beam during search, add the gold
+   * sequence as the last path into the all expanded candidates.
+   */
+  if (goldAsExtraPath_) goldIdsInFinalExpansion_ = pathCount++;
+
+  pathRowIdsInEachBeam_.clear();
+  pathRowIdsInEachBeam_.resize(validExpansionCount_,
+                               std::vector<int>(pathCount, 0));
+  parentIdsInBeam_.clear();
+  parentIdsInBeam_.resize(pathCount, 0);
+
+  if (goldAsExtraPath_) {
+    /* add gold sequence into the total expansion. */
+    pathRowIdsInEachBeam_[beamId].back() =
+        beams_->gold[beamId] +
+        getSeqStartPos(beamId, goldRowIds_[validExpansionCount_ - 1]);
+    parentIdsInBeam_.back() = goldRowIds_[validExpansionCount_ - 1];
+  } else {
+    size_t goldOffset = goldRowIds_[beamId] * beamSize_ + goldColIds_[beamId];
+    goldIdsInFinalExpansion_ =
+        std::count_if(candidates->getData(),
+                      candidates->getData() + goldOffset,
+                      [](const real& val) { return val != -1.; });
+  }
+
+  /*
+   * TODO(caoying): fix this, store the indices of selected candidate
+   * paths into Argument.ids
+   */
+  real* ids = candidates->getData();
+  size_t curIdx = 0;
+  for (size_t i = 0; i < height; ++i) {
+    int basePos = getSeqStartPos(beamId, i);
+    for (size_t j = 0; j < beamSize_; ++j) {
+      int id = ids[i * beamSize_ + j];
+      if (id == -1) continue;
+      pathRowIdsInEachBeam_[beamId][curIdx] = id + basePos;
+      parentIdsInBeam_[curIdx++] = i;
+    }
+  }
+  return pathCount;
+}
+
+void CostForOneSequence::constructTotalExpansion() {
+  /*
+   * construct the entire expanded beam by begining with the last search
+   * in which gold falls off the beam.
+   */
+  size_t totalPathCount = initLastExpansion();
+
+  for (int beamId = validExpansionCount_ - 2; beamId >= 0; --beamId) {
+    const MatrixPtr candidates = beams_->candidateIds[beamId];
+    real* ids = candidates->getData();
+
+    int lastParentIdInBeam = -1;
+    int basePos = -1;
+    for (size_t i = 0;
+         i < (goldAsExtraPath_ ? totalPathCount - 1 : totalPathCount);
+         ++i) {
+      int id = ids[parentIdsInBeam_[i]];
+      int parentRowId = std::div(parentIdsInBeam_[i], beamSize_).quot;
+      if (parentIdsInBeam_[i] != lastParentIdInBeam)
+        basePos = getSeqStartPos(beamId, parentRowId);
+
+      pathRowIdsInEachBeam_[beamId][i] = id + basePos;
+      lastParentIdInBeam = parentIdsInBeam_[i];
+      parentIdsInBeam_[i] = parentRowId;
+
+      if (goldAsExtraPath_)
+        pathRowIdsInEachBeam_[beamId][totalPathCount - 1] =
+            beams_->gold[beamId] + getSeqStartPos(beamId, goldRowIds_[beamId]);
+    }
+  }
+}
+
+real CostForOneSequence::globallyNormalizedScore() {
+  expandedPathScores_.resize(validExpansionCount_);
+
+  Matrix::resizeOrCreate(
+      softmaxOut_, 1, pathRowIdsInEachBeam_[0].size(), false, false);
+  softmaxOut_->zeroMem();
+  MatrixPtr tmp = Matrix::create(
+      softmaxOut_->getData(), softmaxOut_->getWidth(), 1, false, false);
+
+  for (size_t i = 0; i < validExpansionCount_; ++i) {
+    Matrix::resizeOrCreate(expandedPathScores_[i],
+                           pathRowIdsInEachBeam_[i].size(),
+                           1,
+                           false,
+                           false);
+    expandedPathScores_[i]->zeroMem();
+
+    IVectorPtr rowIds = IVector::create(pathRowIdsInEachBeam_[i].data(),
+                                        pathRowIdsInEachBeam_[i].size(),
+                                        false);
+    expandedPathScores_[i]->selectRows(*(beams_->scores[i]), *rowIds);
+    tmp->add(*expandedPathScores_[i]);
+  }
+
+  softmaxOut_->softmax(*softmaxOut_);
+  return -std::log(softmaxOut_->getData()[goldIdsInFinalExpansion_]);
+}
+
+real CostForOneSequence::forward() {
+  calValidExpandStep();
+  constructTotalExpansion();
+  return globallyNormalizedScore();
+}
+
+void CostForOneSequence::backward() {
+  /*
+   * when softmax layer is the output layer, and it is combined with
+   * cross-entropy as cost. The derivate with regard to softmax's input
+   * is simply:
+   *
+   * grad_i = softmax_out_i - target_i,
+   *
+   * and here hard label is used.
+   */
+  softmaxOut_->getData()[goldIdsInFinalExpansion_] -= 1.;
+
+  MatrixPtr tmp = Matrix::create(
+      softmaxOut_->getData(), softmaxOut_->getWidth(), 1, false, false);
+
+  for (size_t i = 0; i < validExpansionCount_; ++i) {
+    IVectorPtr rowIds = IVector::create(pathRowIdsInEachBeam_[i].data(),
+                                        pathRowIdsInEachBeam_[i].size(),
+                                        false);
+    /*
+      beams_->scoreGrad[i] has been intialized outside this class, this
+      class only keeps a pointer pointing to the original input gradients,
+      so here does not need to allocate or initalize the memory.
+    */
+    tmp->addToRows(*beams_->scoreGrad[i], *rowIds);
+  }
+}
+
+REGISTER_LAYER(cross_entropy_over_beam, CrossEntropyOverBeam);
+
+bool CrossEntropyOverBeam::init(const LayerMap& layerMap,
+                                const ParameterMap& parameterMap) {
+  /* Initialize the basic parent class */
+  Layer::init(layerMap, parameterMap);
+  CHECK_EQ(0U, inputLayers_.size() % 3) << "Error input number.";
+
+  beamExpanCount_ = inputLayers_.size() / 3;
+
+  candidateScores_.resize(beamExpanCount_);
+  candidateScoreGrad_.resize(beamExpanCount_);
+
+  candidateInBeam_.resize(beamExpanCount_);
+  goldSequence_.resize(beamExpanCount_);
+  gradToInputs_.resize(beamExpanCount_);
+
+  setNeedSequenceInfo(false);
+  return true;
+}
+
+void CrossEntropyOverBeam::checkInputs() {
+  batchSize_ = 0;
+  for (size_t i = 0; i < beamExpanCount_; ++i) {
+    const Argument& scores = getInput(i * 3);
+    const Argument& selCandidates = getInput(i * 3 + 1);
+    const Argument& goldSeq = getInput(i * 3 + 2);
+
+    if (i) {
+      CHECK(scores.hasSubseq()) << "input " << i << " "
+                                << inputLayers_[i * 3]->getName()
+                                << " should be a nested sequence";
+      CHECK_EQ(getInputValue(i * 3 + 1)->getWidth(), beamSize_);
+      CHECK_EQ(scores.getNumSequences(), batchSize_);
+      CHECK_EQ(scores.getNumSubSequences(), selCandidates.getBatchSize());
+    } else {
+      CHECK(scores.hasSeq()) << "input " << i << " "
+                             << inputLayers_[i]->getName()
+                             << " should be a sequence";
+      batchSize_ = scores.getNumSequences();
+      beamSize_ = getInputValue(i * 3 + 1)->getWidth();
+      CHECK_EQ(batchSize_, selCandidates.getBatchSize());
+    }
+    CHECK_EQ(1U, scores.value->getWidth());
+    CHECK_EQ(batchSize_, goldSeq.getBatchSize());
+  }
+}
+
+void CrossEntropyOverBeam::copyInputsToCpu() {
+  auto copyValue = [](const MatrixPtr& src, MatrixPtr& trg) {
+    if (dynamic_cast<GpuMatrix*>(src.get())) {
+      Matrix::resizeOrCreate(
+          trg, src->getHeight(), src->getWidth(), false, false);
+      trg->copyFrom(*src);
+    } else {
+      trg = std::move(src);
+    }
+  };
+
+  auto copyIds = [](const IVectorPtr& src, IVectorPtr& trg) {
+    if (dynamic_cast<GpuIVector*>(src.get())) {
+      IVector::resizeOrCreate(trg, src->getSize(), false);
+      trg->copyFrom(*src);
+    } else {
+      trg = std::move(src);
+    }
+  };
+
+  beamSplitPos_.clear();
+  beamSplitPos_.resize(batchSize_, std::vector<int>(beamExpanCount_, 0));
+  for (size_t i = 0; i < beamExpanCount_; ++i) {
+    copyValue(getInputValue(i * 3), candidateScores_[i]);
+    copyValue(getInputValue(i * 3 + 1), candidateInBeam_[i]);
+    copyIds(getInput(i * 3 + 2).ids, goldSequence_[i]);
+
+    if (i) {
+      ICpuGpuVectorPtr seqInfo = getInput(i * 3).sequenceStartPositions;
+      const int* seqStarts = seqInfo->getMutableData(false);
+      ICpuGpuVectorPtr subSeqInfo = getInput(i * 3).subSequenceStartPositions;
+      const int* subSeqStarts = subSeqInfo->getMutableData(false);
+
+      size_t seqId = 1;
+      for (size_t subSeqId = 0; subSeqId < subSeqInfo->getSize() - 1;
+           ++subSeqId) {
+        CHECK_LT(seqId, seqInfo->getSize());
+        if (subSeqStarts[subSeqId] == seqStarts[seqId]) {
+          beamSplitPos_[seqId][i] = beamSplitPos_[seqId - 1][i];
+          seqId++;
+        }
+        beamSplitPos_[seqId - 1][i]++;
+      }
+    } else {
+      for (size_t j = 0; j < batchSize_; ++j) beamSplitPos_[j][i] = j + 1;
+    }
+  }
+}
+
+void CrossEntropyOverBeam::splitBatchBeams() {
+  beamCosts_.resize(batchSize_);
+  beamPerSeq_.resize(batchSize_, BeamExpansion(beamExpanCount_));
+
+  for (size_t i = 0; i < beamExpanCount_; ++i) {
+    int* seqStarts =
+        getInput(i * 3).sequenceStartPositions->getMutableData(false);
+
+    int* subSeqStarts = nullptr;
+    int maxLen = 0;
+    if (i) {
+      subSeqStarts =
+          getInput(i * 3).subSequenceStartPositions->getMutableData(false);
+      maxLen = getInput(i * 3).subSequenceStartPositions->getSize() - 1;
+    } else {
+      maxLen = getInput(i).sequenceStartPositions->getSize() - 1;
+    }
+
+    for (size_t j = 0; j < batchSize_; ++j) {
+      beamPerSeq_[j].scores[i] =
+          Matrix::create(candidateScores_[i]->getData() + seqStarts[j],
+                         seqStarts[j + 1] - seqStarts[j],
+                         1,
+                         false,
+                         false);
+      beamPerSeq_[j].scoreGrad[i] =
+          Matrix::create(candidateScoreGrad_[i]->getData() + seqStarts[j],
+                         seqStarts[j + 1] - seqStarts[j],
+                         1,
+                         false,
+                         false);
+
+      int offset = j ? beamSplitPos_[j - 1][i] : 0;
+      int height = beamSplitPos_[j][i] - (j ? beamSplitPos_[j - 1][i] : 0);
+      CHECK_GE(maxLen, offset + height);
+      beamPerSeq_[j].seqInfo[i] = IVector::create(
+          (i ? subSeqStarts : seqStarts) + offset, height + 1, false);
+
+      beamPerSeq_[j].candidateIds[i] =
+          Matrix::create(candidateInBeam_[i]->getData() + offset * beamSize_,
+                         height,
+                         beamSize_,
+                         false,
+                         false);
+      beamPerSeq_[j].gold[i] = goldSequence_[i]->getData()[j];
+
+      CHECK_LE(beamPerSeq_[j].gold[i], seqStarts[j + 1] - seqStarts[j]);
+    }
+  }
+}
+
+void CrossEntropyOverBeam::resizeOutput() {
+  Matrix::resizeOrCreate(output_.value, batchSize_, 1, false, false);
+  output_.value->zeroMem();
+
+  for (size_t i = 0; i < beamExpanCount_; ++i) {
+    MatrixPtr inGrad = getInputGrad(i * 3);
+    if (dynamic_cast<GpuMatrix*>(inGrad.get())) {
+      Matrix::resizeOrCreate(candidateScoreGrad_[i],
+                             inGrad->getHeight(),
+                             inGrad->getWidth(),
+                             false,
+                             false);
+    } else {
+      candidateScoreGrad_[i] = std::move(inGrad);
+    }
+    candidateScoreGrad_[i]->zeroMem();
+  }
+}
+
+void CrossEntropyOverBeam::copyGradToGpu(size_t copyCount) {
+  for (size_t i = 0; i < beamExpanCount_; ++i) {
+    if (dynamic_cast<GpuMatrix*>(getInputGrad(i * 3).get()))
+      getInputGrad(i * 3)->copyFrom(*candidateScoreGrad_[i]);
+
+    if (i == copyCount - 1) break;
+  }
+}
+
+void CrossEntropyOverBeam::forward(PassType passType) {
+  Layer::forward(passType);
+
+  checkInputs();
+  copyInputsToCpu();
+
+  resizeOutput();
+  splitBatchBeams();
+
+  MatrixPtr outputValue = getOutputValue();
+  for (size_t i = 0; i < batchSize_; ++i) {
+    beamCosts_[i].setData(
+        std::move(std::make_shared<BeamExpansion>(beamPerSeq_[i])), beamSize_);
+    outputValue->getData()[i] = beamCosts_[i].forward();
+  }
+}
+
+void CrossEntropyOverBeam::backward(const UpdateCallback& callback) {
+  for (size_t i = 0; i < batchSize_; ++i) {
+    beamCosts_[i].backward();
+    copyGradToGpu(beamCosts_[i].getValidExpansionCount());
+  }
+}
+
+}  // namespace paddle

paddle/gserver/layers/CrossEntropyOverBeam.h

+/* 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. */
+
+#pragma once
+
+#include "CrossEntropyOverBeam.h"
+#include "Layer.h"
+
+namespace paddle {
+
+/* This struct stores the beams in all search steps for a single sequence. */
+struct BeamExpansion {
+  std::vector<MatrixPtr> scores;
+  std::vector<IVectorPtr> seqInfo;
+
+  std::vector<MatrixPtr> candidateIds;
+  std::vector<int> gold;
+
+  std::vector<MatrixPtr> scoreGrad;
+
+  size_t expansionCount;
+
+  explicit BeamExpansion(int n) {
+    expansionCount = n;
+    scores.resize(expansionCount);
+    seqInfo.resize(expansionCount);
+    candidateIds.resize(expansionCount);
+    scoreGrad.resize(expansionCount);
+
+    gold.resize(expansionCount);
+  }
+};
+typedef std::shared_ptr<BeamExpansion> BeamExpansionPtr;
+
+class CostForOneSequence {
+public:
+  CostForOneSequence()
+      : beamSize_(0), validExpansionCount_(0), goldAsExtraPath_(false) {}
+  void setData(const BeamExpansionPtr bPtr, size_t beamSize) {
+    beams_ = bPtr;
+    beamSize_ = beamSize;
+
+    expandedPathScores_.clear();
+    expandedPathScores_.resize(beams_->expansionCount);
+
+    goldRowIds_.clear();
+    goldRowIds_.resize(beams_->expansionCount, 0);
+    goldColIds_.clear();
+    goldColIds_.resize(beams_->expansionCount, -1);
+  }
+  size_t getValidExpansionCount() { return validExpansionCount_; }
+
+  real forward();
+  void backward();
+
+private:
+  void calValidExpandStep();
+  void constructTotalExpansion();
+  size_t initLastExpansion();
+  real globallyNormalizedScore();
+
+  int getSeqStartPos(size_t beamId, size_t rowId) {
+    CHECK_GT(beams_->seqInfo[beamId]->getSize() - 1, rowId);
+    int* starts = beams_->seqInfo[beamId]->getData();
+    return starts[rowId] - starts[0];
+  }
+
+  size_t beamSize_;
+  size_t validExpansionCount_;
+  bool goldAsExtraPath_;
+  std::vector<int> goldRowIds_;
+  std::vector<int> goldColIds_;
+
+  BeamExpansionPtr beams_;
+  std::vector<std::vector<int>> pathRowIdsInEachBeam_;
+  std::vector<int> parentIdsInBeam_;
+  size_t goldIdsInFinalExpansion_;
+
+  std::vector<MatrixPtr> expandedPathScores_;
+
+  MatrixPtr softmaxOut_;
+};
+
+class CrossEntropyOverBeam : public Layer {
+public:
+  explicit CrossEntropyOverBeam(const LayerConfig& config) : Layer(config) {}
+  bool init(const LayerMap& layerMap,
+            const ParameterMap& parameterMap) override;
+  void forward(PassType passType) override;
+  void backward(const UpdateCallback& callback) override;
+
+private:
+  void checkInputs();
+  void copyInputsToCpu();
+  void resizeOutput();
+  void copyGradToGpu(size_t copyCount);
+  void splitBatchBeams();
+
+  size_t beamExpanCount_;
+  size_t batchSize_;
+  size_t beamSize_;
+
+  /*
+   * the process of constructing beams is not friendly to GPU, currently, this
+   * layer only runs on CPU, if any of its inputs is on GPU memory, then copy
+   * it to CPU memory.
+   */
+  std::vector<MatrixPtr> candidateScores_;
+  std::vector<MatrixPtr> candidateScoreGrad_;
+  std::vector<MatrixPtr> candidateInBeam_;
+  std::vector<MatrixPtr> gradToInputs_;
+  std::vector<IVectorPtr> goldSequence_;
+  std::vector<std::vector<int>> beamSplitPos_;
+
+  /*
+   * split entire bath of beams into beam per sequnence and store the result
+   * into this member.
+   */
+  std::vector<BeamExpansion> beamPerSeq_;
+  /* beamCosts_ is used to propagate error in one sequence. */
+  std::vector<CostForOneSequence> beamCosts_;
+};
+
+}  // namespace paddle

paddle/gserver/layers/ExpandConvLayer.cpp

@@ -29,6 +29,10 @@ namespace paddle {
 REGISTER_LAYER(exconv, ExpandConvLayer);
 REGISTER_LAYER(exconvt, ExpandConvLayer);
 
+inline bool isDepthwiseConv(int channels, int groups) {
+  return channels == groups;
+}
+
 bool ExpandConvLayer::init(const LayerMap &layerMap,
                            const ParameterMap &parameterMap) {
   /* Initialize the basic convolutional parent class */
@@ -47,14 +51,27 @@ bool ExpandConvLayer::init(const LayerMap &layerMap,
     std::vector<size_t> paddings = {(size_t)paddingY_[i], (size_t)padding_[i]};
     std::vector<size_t> strides = {(size_t)strideY_[i], (size_t)stride_[i]};
 
-    if (useGpu_ && (size_t)groups_[i] == (size_t)channels_[i] && !isDeconv_) {
+    // Convolution Layer uses the GemmConv function by default.
+    convType = "GemmConv";
+    convGradInputType = "GemmConvGradInput";
+    convGradFilterType = "GemmConvGradFilter";
+
+    // If depth wise convolution and useGpu == true
+    if (useGpu_ && isDepthwiseConv(channels_[i], groups_[i]) && !isDeconv_) {
       convType = "DepthwiseConv";
       convGradInputType = "DepthwiseConvGradInput";
       convGradFilterType = "DepthwiseConvGradFilter";
-    } else {
-      convType = "GemmConv";
-      convGradInputType = "GemmConvGradInput";
-      convGradFilterType = "GemmConvGradFilter";
+    }
+
+    // If depth wise convolution and useGpu == false and ARM-NEON
+    if (!useGpu_ && isDepthwiseConv(channels_[i], groups_[i]) && !isDeconv_) {
+#if defined(__ARM_NEON__) || defined(__ARM_NEON)
+      if ((filterSize_[i] == filterSizeY_[i]) &&
+          (filterSize_[i] == 3 || filterSize_[i] == 4) &&
+          (stride_[i] == strideY_[i]) && (stride_[i] == 1 || stride_[i] == 2)) {
+        convType = "NeonDepthwiseConv";
+      }
+#endif
     }
 
     if (FLAGS_use_nnpack && !isDeconv_) {

paddle/gserver/layers/Layer.cpp

@@ -41,7 +41,7 @@ namespace paddle {
 Layer::Layer(const LayerConfig& config, bool useGpu)
     : config_(config),
       useGpu_(useGpu),
-      deviceId_(-1),
+      deviceId_(CPU_DEVICE),
       needSequenceInfo_(true) {}
 
 bool Layer::init(const LayerMap& layerMap, const ParameterMap& parameterMap) {

paddle/gserver/layers/Layer.h

@@ -59,7 +59,12 @@ protected:
   LayerConfig config_;
   /// whether to use GPU
   bool useGpu_;
-  /// Device Id. CPU is -1, and GPU is 0, 1, 2 ...
+  /// Paddle device ID, MKLDNN is -2, CPU is -1
+  enum PADDLE_DEVICE_ID {
+    MKLDNN_DEVICE = -2,
+    CPU_DEVICE = -1,
+  };
+  /// Device Id. MKLDNN is -2, CPU is -1, and GPU is 0, 1, 2 ...
   int deviceId_;
   /// Input layers
   std::vector<LayerPtr> inputLayers_;
@@ -77,6 +82,7 @@ protected:
   Argument output_;
   /// Several outputs stored on different devices, used in 'parallel_nn' case,
   /// and record them by deviceId_.
+  /// Also used in 'use_mkldnn' case.
   std::vector<Argument> outputOtherDevice_;
   /// If there are several outputs, map them by each name.
   std::map<std::string, Argument*> outputMap_;
@@ -172,6 +178,13 @@ protected:
     return inputLayer.getOutput(deviceId_);
   }
 
+  /**
+   * Get the argument of input layer with deviceId.
+   */
+  const Argument& getInput(size_t inputIndex, int deviceId) const {
+    return inputLayers_[inputIndex]->getOutput(deviceId);
+  }
+
   /**
    * Get the forward-input value.
    */
@@ -186,6 +199,13 @@ protected:
     return inputLayer.getOutput(deviceId_).value;
   }
 
+  /**
+   * Get the forward-input value with deviceId.
+   */
+  const MatrixPtr& getInputValue(int inputIndex, int deviceId) {
+    return inputLayers_[inputIndex]->getOutput(deviceId).value;
+  }
+
   /**
    * Get the forward-input grad.
    */
@@ -200,6 +220,13 @@ protected:
     return inputLayer.getOutput(deviceId_).grad;
   }
 
+  /**
+   * Get the forward-input grad.
+   */
+  const MatrixPtr& getInputGrad(int inputIndex, int deviceId) {
+    return inputLayers_[inputIndex]->getOutput(deviceId).grad;
+  }
+
   /**
    * Get the forward-input label.
    */

paddle/gserver/layers/MKLDNNFcLayer.cpp

@@ -61,43 +61,42 @@ void MKLDNNFcLayer::convertWeightsFromPaddle() {
     return;
   }
 
-  // TODO(TJ): dst format should get from wgtVal_
-  int dstFmt = PARAM_FORMAT_MKLDNN_OI;
-  int srcFmt = weight_->getParameterPtr()->getHeaderFormat();
-  if (srcFmt == dstFmt) {
-    return;
-  }
-
-  // The weight_ is transposed from initial paddle weight
-  MatrixPtr paddleWgt = Matrix::create(
-      weight_->getW()->getData(), iLayerSize_, oc_, false, false);
-
-  // TODO(TJ): remove this print when do not need differ weights
-  std::ostringstream ostr;
-  paddleWgt->print(ostr);
-  VLOG(MKLDNN_ALL) << "Initial Weight from paddle: " << std::endl << ostr.str();
-
-  // The mkldnn weight is transposed from initial paddle matrix
-  MatrixPtr paddleWgtT;
-  paddleWgt->transpose(paddleWgtT, true);
-  weight_->getW()->copyFrom(*paddleWgtT);
-  weight_->getParameterPtr()->setHeaderFormat(dstFmt);
+  CHECK(wgtVal_) << "should have been initialized";
+  bool hasNoSpatial_ = ih_ == 1 && iw_ == 1;
+  auto targetDim = wgtVal_->getDims();
+  auto srcFmt = hasNoSpatial_ ? memory::format::io : memory::format::ihwo;
+  wgtVal_->reorderDataFrom(wgtVal_, srcFmt, targetDim);
   hasInitedWgt_ = true;
 }
 
 void MKLDNNFcLayer::convertWeightsToPaddle() {
-  MatrixPtr dnnWgt = weight_->getW();
-  MatrixPtr paddleWgt;
-  dnnWgt->transpose(paddleWgt, true);
-
-  // copy paddle weight and override on weight_
-  MatrixPtr dnnWgtT = Matrix::create(
-      dnnWgt->getData(), dnnWgt->getWidth(), dnnWgt->getHeight(), false, false);
-  dnnWgtT->copyFrom(*paddleWgt);
+  CHECK(wgtVal_) << "should have been initialized";
+  bool hasNoSpatial_ = ih_ == 1 && iw_ == 1;
+  auto targetDim = wgtVal_->getDims();
+  auto dstFmt = hasNoSpatial_ ? memory::format::io : memory::format::ihwo;
+  wgtVal_->reorderDataTo(wgtVal_, dstFmt, targetDim);
+}
+
+void MKLDNNFcLayer::convertOutputToOtherDevice() {
+  copyOutputInfoToOtherDevice();
+  // find other cpu device and reorder output to cpu device
+  int cnt = 0;
+  for (size_t i = 0; i < outputOtherDevice_.size(); i++) {
+    if (outputOtherDevice_[i].deviceId == CPU_DEVICE) {
+      // fc cpu output value do not need convert
+      // just share point
+      outputOtherDevice_[i].value = output_.value;
+      ++cnt;
+    }
+  }
+
+  if (cnt > 1) {
+    LOG(WARNING) << "should not have more than one CPU devie";
+  }
 }
 
 void MKLDNNFcLayer::reshape() {
-  const Argument& input = getInput(0);
+  const Argument& input = getInput(0, getPrev(0)->getDeviceId());
   int batchSize = input.getBatchSize();
   if (bs_ == batchSize) {
     return;
@@ -111,10 +110,6 @@ void MKLDNNFcLayer::reshape() {
   if (iw_ == 0) {
     iw_ = 1;
   }
-  hasSpatial_ = true;
-  if (ih_ == 1 && iw_ == 1) {
-    hasSpatial_ = false;
-  }
   CHECK_EQ(iLayerSize_, inputLayers_[0]->getSize());
   ic_ = iLayerSize_ / (ih_ * iw_);
   CHECK_EQ(size_t(ic_ * ih_ * iw_), iLayerSize_) << "not divisible";
@@ -135,37 +130,53 @@ void MKLDNNFcLayer::reshape() {
 
 void MKLDNNFcLayer::resetFwd() {
   bool hasBias = biases_ && biases_->getW();
-  real* iData = getInputValue(0)->getData();
-  real* oData = getOutputValue()->getData();
-  real* wData = weight_->getW()->getData();
-  real* bData = hasBias ? biases_->getW()->getData() : NULL;
-
-  // TODO(TJ): below create should be covered in MkldnnMatrix
-  // create memory desc
-  memory::desc iMD = hasSpatial_ ? createMD({bs_, ic_, ih_, iw_}, format::nchw)
-                                 : createMD({bs_, ic_}, format::nc);
-  memory::desc wMD = hasSpatial_ ? createMD({oc_, ic_, ih_, iw_}, format::oihw)
-                                 : createMD({oc_, ic_}, format::oi);
-  memory::desc bMD = bData != NULL ? createMD({oc_}, format::x)
-                                   : createMD({}, format::format_undef);
-  memory::desc oMD = createMD({bs_, oc_}, format::nc);
-
-  // create memory primitive desc and memory self
-  inVal_.reset(new memory(memory::primitive_desc(iMD, engine_), iData));
-  wgtVal_.reset(new memory(memory::primitive_desc(wMD, engine_), wData));
-  outVal_.reset(new memory(memory::primitive_desc(oMD, engine_), oData));
+  const MatrixPtr& wgt = weight_->getW();
+  const MatrixPtr& bias = hasBias ? biases_->getW() : nullptr;
+  const MatrixPtr& out = output_.value;
+
+  if (inputIsOnlyMKLDNN()) {
+    const MatrixPtr& in = getInputValue(0);
+    inVal_ = std::dynamic_pointer_cast<MKLDNNMatrix>(in);
+    CHECK(inVal_) << "Input should be MKLDNNMatrix";
+  } else {
+    CHECK_EQ(getPrev(0)->getDeviceId(), CPU_DEVICE) << "Only support CPU yet";
+    const MatrixPtr& in = getInputValue(0, CPU_DEVICE);
+    inVal_ = MKLDNNMatrix::create(
+        in, memory::dims{bs_, ic_, ih_, iw_}, format::nchw, engine_);
+  }
+  inVal_->downSpatial();
+  wgtVal_ = MKLDNNMatrix::create(
+      wgt, memory::dims{oc_, ic_, ih_, iw_}, format::oihw, engine_);
+  wgtVal_->downSpatial();
+  biasVal_ =
+      hasBias ? MKLDNNMatrix::create(bias, {oc_}, format::x, engine_) : nullptr;
+  outVal_ = MKLDNNMatrix::create(out, {bs_, oc_}, format::nc, engine_);
+
+  // change original output value to mkldnn output value
+  output_.value = std::dynamic_pointer_cast<Matrix>(outVal_);
+  if (!outputIsOnlyMKLDNN()) {
+    convertOutputToOtherDevice();
+  }
 
+  // create forward handle
   prop_kind pk = prop_kind::forward;
-  fc_fwd::desc fwdDesc = bData != NULL ? fc_fwd::desc(pk, iMD, wMD, bMD, oMD)
-                                       : fc_fwd::desc(pk, iMD, wMD, oMD);
+  fc_fwd::desc fwdDesc = hasBias ? fc_fwd::desc(pk,
+                                                inVal_->getMemoryDesc(),
+                                                wgtVal_->getMemoryDesc(),
+                                                biasVal_->getMemoryDesc(),
+                                                outVal_->getMemoryDesc())
+                                 : fc_fwd::desc(pk,
+                                                inVal_->getMemoryDesc(),
+                                                wgtVal_->getMemoryDesc(),
+                                                outVal_->getMemoryDesc());
   fc_fwd::primitive_desc fwdPD = fc_fwd::primitive_desc(fwdDesc, engine_);
-
-  if (bData != NULL) {
-    biasVal_.reset(new memory(memory::primitive_desc(bMD, engine_), bData));
+  if (hasBias) {
     fwd_.reset(new fc_fwd(fwdPD, *inVal_, *wgtVal_, *biasVal_, *outVal_));
   } else {
     fwd_.reset(new fc_fwd(fwdPD, *inVal_, *wgtVal_, *outVal_));
   }
+  printValueFormatFlow();
+
   pipelineFwd_.clear();
   pipelineFwd_.push_back(*fwd_);
 }
@@ -175,45 +186,46 @@ void MKLDNNFcLayer::resetBwd() {
     return;
   }
   needResetBwd_ = false;
-
   bool hasBias = biases_ && biases_->getWGrad();
-  real* iData = getInputValue(0)->getData();
-  real* iDiff = getInputGrad(0) != nullptr ? getInputGrad(0)->getData() : NULL;
-  real* oDiff = getOutputGrad()->getData();
-  real* wDiff = weight_->getWGrad()->getData();
-  real* bDiff = hasBias ? biases_->getWGrad()->getData() : NULL;
 
   /// backward weight
-  // create memory desc for backward memory
-  memory::desc iMD = hasSpatial_ ? createMD({bs_, ic_, ih_, iw_}, format::nchw)
-                                 : createMD({bs_, ic_}, format::nc);
-  memory::desc wMD = hasSpatial_ ? createMD({oc_, ic_, ih_, iw_}, format::oihw)
-                                 : createMD({oc_, ic_}, format::oi);
-  memory::desc oMD = createMD({bs_, oc_}, format::nc);
-  memory::desc bMD = bDiff != NULL ? createMD({oc_}, format::x)
-                                   : createMD({}, format::format_undef);
-
-  if (inVal_) {
-    // update data
-    inVal_->set_data_handle(iData);
-  } else {
-    inVal_.reset(new memory(memory::primitive_desc(iMD, engine_), iData));
-  }
-
-  // create memory primitive desc and memory self
-  wgtGrad_.reset(new memory(memory::primitive_desc(wMD, engine_), wDiff));
-  outGrad_.reset(new memory(memory::primitive_desc(oMD, engine_), oDiff));
-
-  fc_fwd::desc fwdDesc = fc_fwd::desc(prop_kind::forward, iMD, wMD, oMD);
+  CHECK(inVal_) << "Should have input value";
+  const MatrixPtr& wgt = weight_->getWGrad();
+  const MatrixPtr& bias = hasBias ? biases_->getWGrad() : nullptr;
+
+  // TODO(TJ): merge outgrad
+  int device = outputIsOnlyMKLDNN() ? MKLDNN_DEVICE : CPU_DEVICE;
+  // for MKLDNN device:
+  // can not directly cast outputgrad to mkldnnmatrix,
+  // since each layer can not write the inputgrad to mkldnn inputgrad.
+  // So just create from matrix with outputvalue format.
+  // for CPU device:
+  // fc do not need to convert from cpu device since output is always nc format
+  // only need create from cpu device
+  const MatrixPtr& out = getOutput(device).grad;
+  outGrad_ = MKLDNNMatrix::create(out, outVal_->getPrimitiveDesc());
+  wgtGrad_ = MKLDNNMatrix::create(wgt, wgtVal_->getPrimitiveDesc());
+  biasGrad_ = hasBias ? MKLDNNMatrix::create(bias, biasVal_->getPrimitiveDesc())
+                      : nullptr;
+
+  // create memory primitive desc
+  fc_fwd::desc fwdDesc = fc_fwd::desc(prop_kind::forward,
+                                      inVal_->getMemoryDesc(),
+                                      wgtGrad_->getMemoryDesc(),
+                                      outGrad_->getMemoryDesc());
   fc_fwd::primitive_desc fwdPD = fc_fwd::primitive_desc(fwdDesc, engine_);
-  fc_bwdWgt::desc bwdWgtDesc = bDiff != NULL
-                                   ? fc_bwdWgt::desc(iMD, wMD, bMD, oMD)
-                                   : fc_bwdWgt::desc(iMD, wMD, oMD);
+  fc_bwdWgt::desc bwdWgtDesc = hasBias
+                                   ? fc_bwdWgt::desc(inVal_->getMemoryDesc(),
+                                                     wgtGrad_->getMemoryDesc(),
+                                                     biasGrad_->getMemoryDesc(),
+                                                     outGrad_->getMemoryDesc())
+                                   : fc_bwdWgt::desc(inVal_->getMemoryDesc(),
+                                                     wgtGrad_->getMemoryDesc(),
+                                                     outGrad_->getMemoryDesc());
   fc_bwdWgt::primitive_desc bwdWgtPD =
       fc_bwdWgt::primitive_desc(bwdWgtDesc, engine_, fwdPD);
 
-  if (bDiff != NULL) {
-    biasGrad_.reset(new memory(memory::primitive_desc(bMD, engine_), bDiff));
+  if (hasBias) {
     bwdWgt_.reset(
         new fc_bwdWgt(bwdWgtPD, *inVal_, *outGrad_, *wgtGrad_, *biasGrad_));
   } else {
@@ -223,15 +235,26 @@ void MKLDNNFcLayer::resetBwd() {
   pipelineBwd_.push_back(*bwdWgt_);
 
   /// backward data
-  if (iDiff == NULL) {
+  device = inputIsOnlyMKLDNN() ? MKLDNN_DEVICE : CPU_DEVICE;
+  const MatrixPtr& in = getInputGrad(0, device);
+  if (in == nullptr) {
     return;
   }
-  fc_bwdData::desc bwdDataDesc = fc_bwdData::desc(iMD, wMD, oMD);
+  if (getInput(0, device).getAllCount() > 1) {
+    // TODO(TJ): use outputMaps_ ways when merge outgrad done
+  } else {
+    inGrad_ = MKLDNNMatrix::create(in, inVal_->getPrimitiveDesc());
+  }
+
+  fc_bwdData::desc bwdDataDesc = fc_bwdData::desc(inVal_->getMemoryDesc(),
+                                                  wgtGrad_->getMemoryDesc(),
+                                                  outGrad_->getMemoryDesc());
   fc_bwdData::primitive_desc bwdDataPD =
       fc_bwdData::primitive_desc(bwdDataDesc, engine_, fwdPD);
-  inGrad_.reset(new memory(memory::primitive_desc(iMD, engine_), iDiff));
+
   CHECK(wgtVal_) << "Should have weight memory";
   bwdData_.reset(new fc_bwdData(bwdDataPD, *outGrad_, *wgtVal_, *inGrad_));
+  printGradFormatFlow();
   pipelineBwd_.push_back(*bwdData_);
 }
 
@@ -241,11 +264,7 @@ void MKLDNNFcLayer::forward(PassType passType) {
 
   {
     REGISTER_TIMER_INFO("mkldnn_FwdTimer", getName().c_str());
-
-    // update input data
-    // since it might be changed if this is after data layer
-    real* iData = getInputValue(0)->getData();
-    inVal_->set_data_handle(iData);
+    syncInputValue();
 
     // just submit forward pipeline
     stream_->submit(pipelineFwd_);
@@ -267,10 +286,7 @@ void MKLDNNFcLayer::backward(const UpdateCallback& callback) {
     REGISTER_TIMER_INFO("mkldnn_bwdTimer", getName().c_str());
     resetBwd();
 
-    // update diff
-    real* oDiff = getOutputGrad()->getData();
-    outGrad_->set_data_handle(oDiff);
-
+    syncOutputGrad();
     // just sumbmit backward pipeline
     stream_->submit(pipelineBwd_);
   }

paddle/gserver/layers/MKLDNNFcLayer.h

@@ -32,16 +32,13 @@ protected:
   // if has already init the weight
   bool hasInitedWgt_;
 
-  // if input layer has image size info (ih>1 && iw>1)
-  bool hasSpatial_;
-
   // fc weight and bias
   std::unique_ptr<Weight> weight_;
   std::unique_ptr<Weight> biases_;
 
 public:
   explicit MKLDNNFcLayer(const LayerConfig& config)
-      : MKLDNNLayer(config), hasInitedWgt_(false), hasSpatial_(true) {}
+      : MKLDNNLayer(config), hasInitedWgt_(false) {}
 
   ~MKLDNNFcLayer() {}
 
@@ -75,6 +72,8 @@ protected:
    * only would be called when needed
    */
   void resetBwd();
+
+  void convertOutputToOtherDevice() override;
 };
 
 }  // namespace paddle

paddle/gserver/layers/MKLDNNLayer.h

@@ -18,9 +18,9 @@ limitations under the License. */
 #include "Layer.h"
 #include "MKLDNNBase.h"
 #include "mkldnn.hpp"
+#include "paddle/math/MKLDNNMatrix.h"
 
 DECLARE_bool(use_mkldnn);
-DECLARE_bool(use_mkldnn_wgt);
 
 namespace paddle {
 
@@ -52,15 +52,15 @@ protected:
   std::vector<mkldnn::primitive> pipelineFwd_;
   std::vector<mkldnn::primitive> pipelineBwd_;
 
-  // TODO(TJ): change below memory as MKLDNNMatrixPtr type
-  std::shared_ptr<mkldnn::memory> inVal_;
-  std::shared_ptr<mkldnn::memory> inGrad_;
-  std::shared_ptr<mkldnn::memory> outVal_;
-  std::shared_ptr<mkldnn::memory> outGrad_;
-  std::shared_ptr<mkldnn::memory> wgtVal_;
-  std::shared_ptr<mkldnn::memory> wgtGrad_;
-  std::shared_ptr<mkldnn::memory> biasVal_;
-  std::shared_ptr<mkldnn::memory> biasGrad_;
+  // MKLDNNMatrixPtr
+  MKLDNNMatrixPtr inVal_;
+  MKLDNNMatrixPtr inGrad_;
+  MKLDNNMatrixPtr outVal_;
+  MKLDNNMatrixPtr outGrad_;
+  MKLDNNMatrixPtr wgtVal_;
+  MKLDNNMatrixPtr wgtGrad_;
+  MKLDNNMatrixPtr biasVal_;
+  MKLDNNMatrixPtr biasGrad_;
 
 public:
   explicit MKLDNNLayer(const LayerConfig& config)
@@ -83,17 +83,21 @@ public:
 
   virtual bool init(const LayerMap& layerMap,
                     const ParameterMap& parameterMap) {
+    CHECK(FLAGS_use_mkldnn) << "MkldnnLayers only support use_mkldnn."
+                            << "Please set WITH_MKLDNN=ON "
+                            << "and set use_mkldnn=True";
+    CHECK(!useGpu_) << "Do not support GPU yet";
+
+    // set device id before Layer::init
+    setDevice(MKLDNN_DEVICE);
+    // change param device to MKLDNN device
+    setParamsDevice(MKLDNN_DEVICE, parameterMap);
     if (!Layer::init(layerMap, parameterMap)) {
       return false;
     }
 
-    CHECK(FLAGS_use_mkldnn) << "MkldnnLayers only support use_mkldnn."
-                            << "Please set WITH_MKLDNN=ON "
-                            << "and set use_mkldnn=True";
     stream_.reset(new MKLDNNStream());
     engine_ = CPUEngine::Instance().getEngine();
-
-    // TODO(TJ): deivecId
     return true;
   }
 
@@ -109,6 +113,12 @@ public:
    */
   virtual void convertWeightsToPaddle() {}
 
+  /**
+   * convert MKLDNN output to other device.
+   * only support CPU device yet
+   */
+  virtual void convertOutputToOtherDevice() {}
+
   /**
    * print info about sizes
    */
@@ -118,14 +128,124 @@ public:
                        << ", oh: " << oh_ << ", ow: " << ow_;
   }
 
-  // TODO(TJ): move to MkldnnMatrix
-  // create memory desc
-  inline mkldnn::memory::desc createMD(
-      mkldnn::memory::dims dims,
-      mkldnn::memory::format fmt,
-      mkldnn::memory::data_type type = mkldnn::memory::data_type::f32) {
-    // TODO(TJ): isFmtSuppoted(fmt)
-    return mkldnn::memory::desc(dims, type, fmt);
+  /**
+   * Print the mkldnn memory format flow of value
+   */
+  virtual void printValueFormatFlow() {
+    if (inVal_ && outVal_) {
+      VLOG(MKLDNN_FMTS) << "value format flow --- " << inVal_->getFormat()
+                        << " >>> " << outVal_->getFormat();
+    }
+  }
+
+  /**
+   * Print the mkldnn memory format flow of grad
+   */
+  virtual void printGradFormatFlow() {
+    if (inGrad_ && outGrad_) {
+      VLOG(MKLDNN_FMTS) << "grad format flow --- " << inGrad_->getFormat()
+                        << " <<< " << outGrad_->getFormat();
+    }
+  }
+
+protected:
+  /**
+   * copy image size and sequence info to other device
+   * @note: can not directly use Layer::copyOutputToOtherDevice since here only
+   *        copy base info and do not copy data value
+   */
+  void copyOutputInfoToOtherDevice() {
+    for (size_t i = 0; i < outputOtherDevice_.size(); i++) {
+      outputOtherDevice_[i].setFrameHeight(output_.getFrameHeight());
+      outputOtherDevice_[i].setFrameWidth(output_.getFrameWidth());
+      outputOtherDevice_[i].sequenceStartPositions =
+          output_.sequenceStartPositions;
+      outputOtherDevice_[i].subSequenceStartPositions =
+          output_.subSequenceStartPositions;
+      outputOtherDevice_[i].cpuSequenceDims = output_.cpuSequenceDims;
+    }
+  }
+
+  /**
+   * If input only has MKLDNN device.
+   * Otherwise, only support the previous layer using CPU device.
+   */
+  bool inputIsOnlyMKLDNN(int index = 0) {
+    int prevDevice = getPrev(index)->getDeviceId();
+    if (prevDevice == MKLDNN_DEVICE) {
+      return true;
+    } else {
+      // do not support GPU yet
+      CHECK_EQ(prevDevice, CPU_DEVICE) << "Only support CPU yet";
+      return false;
+    }
+  }
+
+  /**
+   * If output only has MKLDNN device.
+   * Otherwise, other devices should only using CPU device.
+   */
+  bool outputIsOnlyMKLDNN() {
+    for (size_t i = 0; i < outputOtherDevice_.size(); i++) {
+      CHECK_EQ(outputOtherDevice_[i].deviceId, CPU_DEVICE)
+          << "Only support other device is CPU yet";
+    }
+    return outputOtherDevice_.size() == 0;
+  }
+
+  /**
+   * Sync input value data
+   */
+  void syncInputValue() {
+    if (inputIsOnlyMKLDNN()) {
+      return;
+    }
+    real* iData = getInputValue(0, CPU_DEVICE)->getData();
+    // update input data
+    // since it might be changed if this is after data layer
+    inVal_->updateData(iData);
+  }
+
+  /**
+   * Sync output grad data
+   */
+  void syncOutputGrad() {
+    if (outputIsOnlyMKLDNN()) {
+      return;
+    }
+
+    // update diff
+    real* oDiff = getOutput(CPU_DEVICE).grad->getData();
+    outGrad_->updateData(oDiff);
+  }
+
+  /**
+   * Set deviceId of this layer.
+   */
+  void setDevice(int id) { deviceId_ = id; }
+
+  /**
+   * Set deviceId of the params used in this layer.
+   */
+  void setParamsDevice(int id, const ParameterMap& parameterMap) {
+    for (auto& inputConfig : config_.inputs()) {
+      if (inputConfig.has_input_parameter_name()) {
+        ParameterPtr parameter;
+        std::string name = inputConfig.input_parameter_name();
+        CHECK(mapGet(name, parameterMap, &parameter))
+            << "Cannot find input parameter " << name << " for layer "
+            << getName();
+        parameter->setDevice(id);
+      }
+    }
+    if (config_.has_bias_parameter_name()) {
+      ParameterPtr parameter;
+      std::string name = config_.bias_parameter_name();
+      CHECK(mapGet(name, parameterMap, &parameter))
+          << "Cannot find bias parameter " << name << " for layer "
+          << getName();
+      parameter->setDevice(id);
+    }
   }
 };
 

paddle/gserver/tests/CMakeLists.txt

@@ -34,6 +34,13 @@ add_unittest_without_exec(test_CRFLayerGrad
 add_test(NAME test_CRFLayerGrad
     COMMAND test_CRFLayerGrad)
 
+################ test_CrossEntropyOverBeam ####################
+add_unittest_without_exec(test_CrossEntropyOverBeam
+    test_CrossEntropyOverBeamGrad.cpp
+    LayerGradUtil.cpp)
+add_test(NAME test_CrossEntropyOverBeam
+    COMMAND test_CrossEntropyOverBeam)
+
 ################ test_SeqSliceLayerGrad ####################
 add_unittest_without_exec(test_SeqSliceLayerGrad
     test_SeqSliceLayerGrad.cpp

paddle/gserver/tests/test_CrossEntropyOverBeamGrad.cpp

+/* Copyright (c) 2016 Baidu, Inc. 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. */
+
+#include <random>
+#include <sstream>
+
+#include <gtest/gtest.h>
+#include "ModelConfig.pb.h"
+#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/trainer/Trainer.h"
+
+#include "LayerGradUtil.h"
+#include "paddle/testing/TestUtil.h"
+
+using namespace paddle;  // NOLINT
+
+DECLARE_int32(gpu_id);
+DECLARE_bool(thread_local_rand_use_global_seed);
+
+const size_t MAX_SEQ_NUM = 23;
+const size_t MAX_SEQ_LEN = 50;
+const size_t MAX_BEAM_SIZE = 27;
+
+const size_t SEED = (size_t)(time(NULL));
+
+struct SingleBeamExpansion {
+  vector<int> seqStartPos;
+  vector<int> subSeqStartPos;
+  vector<real> candidateScores;
+
+  // TODO(caoying): store this into Argument.ids
+  vector<real> selectedIndices;
+
+  vector<int> groundTruth;
+  vector<size_t> inBeam;
+  vector<int> rowIdxInBeam;
+  vector<int> colIdxInBeam;
+
+  void resetGroundTruth(size_t n) {
+    groundTruth.clear();
+    groundTruth.resize(n, -1);
+
+    inBeam.clear();
+    inBeam.resize(n, 0);
+
+    rowIdxInBeam.clear();
+    rowIdxInBeam.resize(n, -1);
+
+    colIdxInBeam.clear();
+    colIdxInBeam.resize(n, -1);
+  }
+};
+
+inline float randFloat() {
+  return static_cast<float>(rand()) / static_cast<float>(RAND_MAX);
+}
+
+void genRand(real* numbers, size_t n) {
+  default_random_engine generator;
+  uniform_real_distribution<real> distribution(0.0, 1.0);
+  for (size_t i = 0; i < n; ++i) numbers[i] = distribution(generator);
+}
+
+vector<real> randSampling(real range, int n) {
+  CHECK_GE(range, n);
+  vector<real> num(range);
+  iota(begin(num), end(num), 0.);
+  if (range == n) return num;
+
+  random_shuffle(begin(num), end(num));
+  num.resize(n);
+  sort(begin(num), end(num));
+  return num;
+}
+
+void genCandidateScores(bool hasSubseq,
+                        size_t beamSize,
+                        SingleBeamExpansion& prevBeam,
+                        SingleBeamExpansion& curBeam) {
+  vector<int>& seqStartPos = curBeam.seqStartPos;
+  seqStartPos.resize(1, 0);
+  vector<int>& subSeqStartPos = curBeam.subSeqStartPos;
+  subSeqStartPos.resize(1, 0);
+
+  srand(SEED);
+  if (prevBeam.selectedIndices.size()) {
+    if (prevBeam.subSeqStartPos.size() > 1) {
+      int seqIdx = 1;
+      // samples in previous beam are nested sequences.
+      for (size_t i = 1; i < prevBeam.subSeqStartPos.size(); ++i) {
+        for (size_t j = 0; j < beamSize; ++j) {
+          if (prevBeam.selectedIndices[(i - 1) * beamSize + j] == -1.) break;
+          subSeqStartPos.push_back(1 + (rand() % MAX_SEQ_LEN) +
+                                   subSeqStartPos.back());
+        }
+        if (prevBeam.seqStartPos[seqIdx] == prevBeam.subSeqStartPos[i]) {
+          seqStartPos.push_back(subSeqStartPos.back());
+          seqIdx++;
+        }
+      }
+    } else {
+      for (size_t i = 0; i <= prevBeam.selectedIndices.size(); ++i) {
+        if (i && i % beamSize == 0) {
+          seqStartPos.push_back(subSeqStartPos.back());
+          if (i == prevBeam.selectedIndices.size()) break;
+        }
+        if (prevBeam.selectedIndices[i] == -1.) continue;
+        subSeqStartPos.push_back(subSeqStartPos.back() +
+                                 (1 + (rand() % MAX_SEQ_LEN)));
+      }
+    }
+  } else {
+    // the first beam expansion
+    int seqNum = 1 + (rand() % MAX_SEQ_NUM);
+    for (int i = 0; i < seqNum; ++i) {
+      if (hasSubseq) {
+        for (size_t j = 0; j < 1 + (rand() % MAX_SEQ_NUM); ++j)
+          subSeqStartPos.push_back(subSeqStartPos.back() +
+                                   (1 + (rand() % MAX_SEQ_LEN)));
+        seqStartPos.push_back(subSeqStartPos.back());
+      } else {
+        seqStartPos.push_back(seqStartPos.back() +
+                              (1 + (rand() % MAX_SEQ_LEN)));
+      }
+    }
+  }
+
+  size_t totalSeqNum = hasSubseq ? subSeqStartPos.back() : seqStartPos.back();
+  curBeam.candidateScores.resize(totalSeqNum, 0.);
+  genRand(curBeam.candidateScores.data(), totalSeqNum);
+}
+
+void genSelectedIndices(size_t beamSize,
+                        vector<int>& seqStartPos,
+                        vector<real>& selectedIndices) {
+  size_t selectedIdsCount = beamSize * (seqStartPos.size() - 1);
+  selectedIndices.resize(selectedIdsCount, -1.);
+
+  for (size_t i = 0; i < seqStartPos.size() - 1; ++i) {
+    int seqLen = seqStartPos[i + 1] - seqStartPos[i];
+    int n = min(seqLen, static_cast<int>(beamSize));
+    vector<real> ids = randSampling(seqLen, n);
+    memcpy(selectedIndices.data() + i * beamSize,
+           ids.data(),
+           sizeof(real) * ids.size());
+  }
+}
+
+void genGroundTruth(vector<SingleBeamExpansion>& beamExpansions,
+                    size_t beamSize) {
+  SingleBeamExpansion& beam = beamExpansions[1];
+  size_t seqNum = beam.seqStartPos.size() - 1;
+  for (size_t i = 2; i < beamExpansions.size(); ++i)
+    CHECK_EQ(seqNum, beamExpansions[i].seqStartPos.size() - 1);
+
+  srand(SEED);
+
+  // initialize the first beam.
+  beam.resetGroundTruth(seqNum);
+  for (size_t i = 0; i < seqNum; ++i) {
+    if (randFloat() > 0.5) {
+      /*
+       * force the randomly generated label falls in the beam by chance 0.5.
+       * otherwise, when sequence length is relatively long and beam size is
+       * relatively small, the gold sequences falls off the beam at in the
+       * first search.
+       */
+      real* begPos = beam.selectedIndices.data() + i * beamSize;
+      beam.colIdxInBeam[i] =
+          rand() % count_if(begPos, begPos + beamSize, [](const real& val) {
+            return val != -1.;
+          });
+      beam.groundTruth[i] =
+          beam.selectedIndices[i * beamSize + beam.colIdxInBeam[i]];
+      beam.inBeam[i] = 1;
+    } else {
+      int label = rand() % (beam.seqStartPos[i + 1] - beam.seqStartPos[i]);
+      beam.groundTruth[i] = label;
+
+      real* begPos = beam.selectedIndices.data() + i * beamSize;
+      real* endPos = begPos + beamSize;
+      real* lblPos = find(begPos, endPos, real(label));
+      if (lblPos != endPos) {
+        beam.inBeam[i] = 1;
+        beam.colIdxInBeam[i] = lblPos - begPos;
+      }
+    }
+    beam.rowIdxInBeam[i] = i;
+  }
+
+  // iterate over each beam expansions
+  for (size_t i = 2; i < beamExpansions.size(); ++i) {
+    SingleBeamExpansion& curBeam = beamExpansions[i];
+    SingleBeamExpansion& prevBeam = beamExpansions[i - 1];
+    curBeam.resetGroundTruth(seqNum);
+
+    // iterate over each sequence
+    for (size_t j = 0; j < seqNum; ++j) {
+      if (!prevBeam.inBeam[j]) continue;
+
+      // gold sequence falls in the beam in previous search.
+      real* begPos = prevBeam.selectedIndices.data();
+      int offset =
+          prevBeam.rowIdxInBeam[j] * beamSize + prevBeam.colIdxInBeam[j];
+      curBeam.rowIdxInBeam[j] = count_if(
+          begPos, begPos + offset, [](const real& val) { return val != -1.; });
+
+      if (randFloat() > 0.5) {
+        // force the randomly generated label falls in the beam by chance 0.5.
+
+        real* start =
+            curBeam.selectedIndices.data() + curBeam.rowIdxInBeam[j] * beamSize;
+        int n = rand() % count_if(start, start + beamSize, [](const real& val) {
+                  return val != -1.;
+                });
+        curBeam.colIdxInBeam[j] = n;
+        curBeam.groundTruth[j] = *(start + n);
+        curBeam.inBeam[j] = 1;
+      } else {
+        CHECK_LE(curBeam.rowIdxInBeam[j] + 1,
+                 curBeam.subSeqStartPos.size() - 1);
+        int start = curBeam.subSeqStartPos[curBeam.rowIdxInBeam[j]];
+        int end = curBeam.subSeqStartPos[curBeam.rowIdxInBeam[j] + 1];
+        CHECK_GT(size_t(end), size_t(start));
+        int label = rand() % (end - start);
+
+        curBeam.groundTruth[j] = label;
+        real* findBeg =
+            curBeam.selectedIndices.data() + curBeam.rowIdxInBeam[j] * beamSize;
+        real* lblPos =
+            find(findBeg, findBeg + beamSize, static_cast<real>(label));
+        if (lblPos != (findBeg + beamSize)) {
+          curBeam.inBeam[j] = 1;
+          curBeam.colIdxInBeam[j] = lblPos - findBeg;
+        }
+      }
+    }
+  }
+}
+
+void genOneBeam(size_t beamSize,
+                bool hasSubseq,
+                SingleBeamExpansion& prevBeam,
+                SingleBeamExpansion& curBeam) {
+  genCandidateScores(hasSubseq, beamSize, prevBeam, curBeam);
+  genSelectedIndices(beamSize,
+                     hasSubseq ? curBeam.subSeqStartPos : curBeam.seqStartPos,
+                     curBeam.selectedIndices);
+}
+
+void genRandomBeamExpansion(size_t expansionCount,
+                            size_t beamSize,
+                            vector<SingleBeamExpansion>& beamExpansions) {
+  beamExpansions.clear();
+  beamExpansions.resize(expansionCount + 1);
+
+  // beamExpansions[0] is reserved.
+  for (size_t i = 1; i <= expansionCount; ++i)
+    genOneBeam(beamSize, bool(i - 1), beamExpansions[i - 1], beamExpansions[i]);
+  genGroundTruth(beamExpansions, beamSize);
+}
+
+void testCrossEntropyOverBeam(bool useGpu,
+                              size_t beamSize,
+                              vector<SingleBeamExpansion>& beams) {
+  TestConfig config;
+  config.layerConfig.set_type("cross_entropy_over_beam");
+
+  size_t seqNum = 0;
+  for (size_t i = 1; i < beams.size(); ++i) {
+    const SingleBeamExpansion& beam = beams[i];
+    // create scores for all the candidates
+    MatrixPtr candidateScorePtr =
+        Matrix::create(beam.candidateScores.size(), 1, false, false);
+    candidateScorePtr->copyFrom(beam.candidateScores.data(),
+                                beam.candidateScores.size());
+
+    ostringstream paramName;
+    paramName << "candidate_scores_" << i;
+
+    if (beam.subSeqStartPos.size() > 1) {
+      seqNum = beam.subSeqStartPos.size() - 1;
+      config.inputDefs.push_back({INPUT_SELF_DEFINE_DATA,
+                                  paramName.str(),
+                                  candidateScorePtr,
+                                  beam.seqStartPos,
+                                  beam.subSeqStartPos});
+    } else {
+      seqNum = beam.seqStartPos.size() - 1;
+      config.inputDefs.push_back({INPUT_SELF_DEFINE_DATA,
+                                  paramName.str(),
+                                  candidateScorePtr,
+                                  beam.seqStartPos});
+    }
+    config.layerConfig.add_inputs();
+
+    // create indices for the selected candidates
+    MatrixPtr selectedCandidates =
+        Matrix::create(seqNum, beamSize, false, false);
+    selectedCandidates->copyFrom(beam.selectedIndices.data(),
+                                 beam.selectedIndices.size());
+    paramName.clear();
+    paramName << "selected_candidates_" << i;
+    config.inputDefs.push_back(
+        {INPUT_SELF_DEFINE_DATA, paramName.str(), selectedCandidates});
+    config.layerConfig.add_inputs();
+
+    // create the ground truth
+    paramName.clear();
+    paramName << "label_" << i;
+    config.inputDefs.push_back(
+        {INPUT_SELF_DEFINE_DATA, paramName.str(), beam.groundTruth});
+    config.layerConfig.add_inputs();
+  }
+
+  testLayerGrad(
+      config, "cross_entropy_over_beam", seqNum, false, useGpu, false);
+}
+
+TEST(Layer, CrossEntropyOverBeam) {
+  LOG(INFO) << "SEED = " << SEED;
+  const size_t beamSize = 1 + rand() % MAX_BEAM_SIZE;
+  LOG(INFO) << "beamSize = " << beamSize;
+
+  // TODO(caoying): test with random beam expansions.
+  const size_t expansionCount = 3;
+  vector<SingleBeamExpansion> beams;
+  genRandomBeamExpansion(expansionCount, beamSize, beams);
+
+  for (bool useGpu : {false, true})
+    testCrossEntropyOverBeam(useGpu, beamSize, beams);
+}
+
+int main(int argc, char** argv) {
+  initMain(argc, argv);
+  hl_start();
+  hl_init(FLAGS_gpu_id);
+  FLAGS_thread_local_rand_use_global_seed = true;
+  srand(SEED);
+  testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

paddle/math/Allocator.h

@@ -48,7 +48,13 @@ public:
    */
   virtual void* alloc(size_t size) {
     void* ptr;
+#ifdef PADDLE_USE_MKLDNN
+    // refer to https://github.com/01org/mkl-dnn/blob/master/include/mkldnn.hpp
+    // memory alignment
+    CHECK_EQ(posix_memalign(&ptr, 4096ul, size), 0);
+#else
     CHECK_EQ(posix_memalign(&ptr, 32ul, size), 0);
+#endif
     CHECK(ptr) << "Fail to allocate CPU memory: size=" << size;
     return ptr;
   }

paddle/math/CMakeLists.txt

@@ -14,6 +14,17 @@
 #
 file(GLOB MATH_HEADERS . *.h)
 file(GLOB MATH_SOURCES . *.cpp)
+
+if(NOT WITH_MKLDNN)
+    set(DNN_HEADER "${CMAKE_CURRENT_SOURCE_DIR}/MKLDNNMatrix.h")
+    set(DNN_SOURCE "${CMAKE_CURRENT_SOURCE_DIR}/MKLDNNMatrix.cpp")
+    list(REMOVE_ITEM MATH_HEADERS "${DNN_HEADER}")
+    list(REMOVE_ITEM MATH_SOURCES "${DNN_SOURCE}")
+    message(STATUS "Skip compiling with MKLDNNMatrix")
+else()
+    message(STATUS "Compile with MKLDNNMatrix")
+endif()
+
 set(MATH_SOURCES
     "${PADDLE_SOURCE_DIR}/paddle/math/BaseMatrix.cu"
     "${PADDLE_SOURCE_DIR}/paddle/math/TrainingAlgorithmOp.cu"

paddle/math/MKLDNNMatrix.cpp

+/* Copyright (c) 2017 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. */
+
+#include "MKLDNNMatrix.h"
+
+using namespace mkldnn;  // NOLINT
+
+namespace paddle {
+
+MKLDNNMatrixPtr MKLDNNMatrix::create(MatrixPtr m, memory::primitive_desc pd) {
+  memory::desc md = pd.desc();
+  size_t ndims = md.data.ndims;
+  int* dims = md.data.dims;
+  CHECK(ndims > 0) << "Input dims should not be empty";
+  size_t cnts = 1;
+  for (size_t i = 0; i < ndims; ++i) {
+    cnts *= dims[i];
+  }
+
+  if (m == nullptr) {
+    size_t height = dims[0];
+    size_t width = cnts / dims[0];
+    m = Matrix::create(height, width, false, false);
+  }
+
+  CHECK(m) << " Matrix should not be empty";
+  CpuMatrixPtr cpuMatrix = std::dynamic_pointer_cast<CpuMatrix>(m);
+  CHECK(cpuMatrix) << "Only support create from CPU matrix yet";
+
+  CHECK_EQ(cnts, m->getElementCnt()) << "Count size does not match";
+  return std::make_shared<MKLDNNMatrix>(
+      m->getData(), m->getHeight(), m->getWidth(), pd);
+}
+
+MKLDNNMatrixPtr MKLDNNMatrix::create(MatrixPtr m,
+                                     memory::dims dims,
+                                     memory::format fmt,
+                                     engine& eg,
+                                     mkldnn::memory::data_type dtype) {
+  return create(m, memory::primitive_desc(memory::desc(dims, dtype, fmt), eg));
+}
+
+void MKLDNNMatrix::reorderDataFrom(const MKLDNNMatrixPtr& m,
+                                   memory::format srcFmt,
+                                   memory::dims targetDim) {
+  memory::format dstFmt = getFormat();
+  if (srcFmt == dstFmt) {
+    return;
+  }
+  CHECK_EQ(getElementCnt(), m->getElementCnt()) << "size should equal";
+  reorderOnce(getData(), m->getData(), srcFmt, dstFmt, targetDim);
+}
+
+void MKLDNNMatrix::reorderDataTo(const MKLDNNMatrixPtr& m,
+                                 memory::format dstFmt,
+                                 memory::dims targetDim) {
+  memory::format srcFmt = getFormat();
+  if (srcFmt == dstFmt) {
+    return;
+  }
+  CHECK_EQ(getElementCnt(), m->getElementCnt()) << "size should equal";
+  reorderOnce(getData(), m->getData(), srcFmt, dstFmt, targetDim);
+}
+
+void MKLDNNMatrix::reorderOnce(void* srcData,
+                               void* dstData,
+                               memory::format srcFmt,
+                               memory::format dstFmt,
+                               memory::dims dm) {
+  CHECK(srcData);
+  CHECK(dstData);
+  MatrixPtr tmpSrc;
+  if (dstData == srcData) {
+    // inplace data
+    size_t sz = 1;
+    for (size_t i = 0; i < dm.size(); ++i) {
+      sz *= dm[i];
+    }
+    tmpSrc = Matrix::create(sz, 1, false, false);
+    tmpSrc->copyFrom((real*)srcData, sz);
+    srcData = tmpSrc->getData();
+  }
+
+  auto dtype = this->getDtype();
+  auto srcMD = memory::desc(dm, dtype, srcFmt);
+  auto dstMD = memory::desc(dm, dtype, dstFmt);
+
+  auto eg = this->getEngine();
+  auto src = memory(memory::primitive_desc(srcMD, eg), srcData);
+  auto dst = memory(memory::primitive_desc(dstMD, eg), dstData);
+
+  auto r = reorder(src, dst);
+  stream(stream::kind::eager).submit({r}).wait();
+}
+
+void MKLDNNMatrix::downSpatial() {
+  int fmt = getFormat();
+  if (!(fmt == memory::format::nchw || fmt == memory::format::oihw)) {
+    // only support nchw and oihw yet, later can support more like nhwc, ihwo
+    return;
+  }
+
+  // TODO(TJ): change H(height) and W(width) if support nhwc or more
+  const int H = 2, W = 3;
+  memory::dims srcDims = getDims();
+  if (srcDims[H] != 1 || srcDims[W] != 1) {
+    // can not down spatial
+    return;
+  }
+
+  memory::dims dstDims = memory::dims{srcDims[0], srcDims[1]};
+  memory::format dstFmt;
+  switch (fmt) {
+    case memory::format::nchw:
+      dstFmt = memory::format::nc;
+      break;
+    case memory::format::oihw:
+      dstFmt = memory::format::oi;
+      break;
+    default:
+      LOG(FATAL) << "unsupported format";
+  }
+  memory::desc md = memory::desc(dstDims, getDtype(), dstFmt);
+  memory::primitive_desc pd = memory::primitive_desc(md, getEngine());
+  mkldnn_primitive_t result;
+  mkldnn::error::wrap_c_api(
+      mkldnn_primitive_create(&result, pd.get(), nullptr, nullptr),
+      "could not create a memory primitive");
+  reset(result);
+  set_data_handle(getData());
+}
+
+}  // namespace paddle

paddle/math/MKLDNNMatrix.h

+/* Copyright (c) 2017 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. */
+
+#pragma once
+
+#include <vector>
+#include "Matrix.h"
+#include "mkldnn.hpp"
+#include "paddle/parameter/Parameter.h"
+
+namespace paddle {
+
+class MKLDNNMatrix;
+typedef std::shared_ptr<MKLDNNMatrix> MKLDNNMatrixPtr;
+
+/**
+ * @brief MKLDNN Matrix.
+ *
+ */
+class MKLDNNMatrix : public CpuMatrix, public mkldnn::memory {
+public:
+  MKLDNNMatrix(real* data,
+               size_t height,
+               size_t width,
+               mkldnn::memory::primitive_desc pd)
+      : CpuMatrix(data, height, width, false), mkldnn::memory(pd, data) {}
+
+  ~MKLDNNMatrix() {}
+
+  /**
+   * Create MKLDNNMatrix from a MatrixPtr and memory primitive_desc
+   */
+  static MKLDNNMatrixPtr create(MatrixPtr m, mkldnn::memory::primitive_desc pd);
+
+  /**
+   * Create MKLDNNMatrix from a MatrixPtr and memory details info
+   */
+  static MKLDNNMatrixPtr create(
+      MatrixPtr m,
+      mkldnn::memory::dims dims,
+      mkldnn::memory::format fmt,
+      mkldnn::engine& eg,
+      mkldnn::memory::data_type dtype = mkldnn::memory::data_type::f32);
+
+public:
+  /**
+   * Reorder this MKLDNNMatrix from other format.
+   * Support inplace reorder.
+   * @note: this function would only reorder the data layout.
+   *        will NOT change this original dim or format info
+   */
+  void reorderDataFrom(const MKLDNNMatrixPtr& m,
+                       memory::format srcFmt,
+                       memory::dims targetDim);
+
+  /**
+   * Reorder this MKLDNNMatrix to other format.
+   * Support inplace reorder.
+   * @note: this function would only reorder the data layout.
+   *        will NOT change the dst dim or format info
+   */
+  void reorderDataTo(const MKLDNNMatrixPtr& m,
+                     memory::format dstFmt,
+                     memory::dims targetDim);
+
+  /**
+   * Dimensionality reduction.
+   * Change format "nchw --> nc" or "oihw --> oi" if the h and w are both 1
+   */
+  void downSpatial();
+
+  /**
+   * Update the memory data handle.
+   * Caution: This will not check the buffer size of the data,
+   *          it should be coverd by user.
+   */
+  void updateData(void* data) { set_data_handle(data); }
+
+  /**
+   * Get primitive descriptor.
+   */
+  mkldnn::memory::primitive_desc getPrimitiveDesc() {
+    return this->get_primitive_desc();
+  }
+
+  /**
+   * Get memory descriptor.
+   */
+  mkldnn::memory::desc getMemoryDesc() { return getPrimitiveDesc().desc(); }
+
+  /**
+   * Get dimensions.
+   */
+  mkldnn::memory::dims getDims() {
+    mkldnn::memory::desc md = getMemoryDesc();
+    const int* src = md.data.dims;
+    int ndims = md.data.ndims;
+    mkldnn::memory::dims dst;
+    dst.resize(ndims);
+    for (int i = 0; i < ndims; ++i) {
+      dst[i] = src[i];
+    }
+    return dst;
+  }
+
+  /**
+   * Get format.
+   */
+  mkldnn::memory::format getFormat() {
+    return (mkldnn::memory::format)(getMemoryDesc().data.format);
+  }
+
+  /**
+   * Get memory data type.
+   */
+  mkldnn::memory::data_type getDtype() {
+    return (mkldnn::memory::data_type)(getMemoryDesc().data.data_type);
+  }
+
+  /**
+   * Get engine.
+   */
+  mkldnn::engine getEngine() { return getPrimitiveDesc().get_engine(); }
+
+protected:
+  /**
+   * Do reorder once.
+   * Can support inplace.
+   */
+  void reorderOnce(void* srcData,
+                   void* dstData,
+                   memory::format srcFmt,
+                   memory::format dstFmt,
+                   memory::dims dm);
+};
+
+}  // namespace paddle

paddle/parameter/Argument.cpp

@@ -679,6 +679,7 @@ void Argument::reorganizeSeqInfo(
     const ICpuGpuVectorPtr subSeqStartPos,
     std::vector<std::vector<int>>& reorganizedSeqInfo) {
   CHECK(seqStartPos);
+  reorganizedSeqInfo.clear();
 
   int seqNum = seqStartPos->getSize() - 1;
   int* seqStarts = seqStartPos->getMutableData(false);

paddle/parameter/Parameter.h

@@ -281,7 +281,11 @@ public:
   /**
    * @brief Set the format in header.
    */
-  void setHeaderFormat(int32_t fmt) { headerFormat_ = fmt; }
+  void setHeaderFormat(int32_t fmt) {
+    CHECK(isHeaderFormatSupported(fmt)) << "Unsupported format version: "
+                                        << fmt;
+    headerFormat_ = fmt;
+  }
 
   /**
    * @brief  Parameter Update Hook.

paddle/pserver/LightNetwork.cpp

@@ -22,7 +22,6 @@ limitations under the License. */
 
 #include <arpa/inet.h>
 #include <net/if.h>
-#include <net/if_arp.h>
 #include <sys/ioctl.h>
 #include <sstream>
 

python/paddle/trainer/config_parser.py

@@ -1698,6 +1698,21 @@ class MultiClassCrossEntropySelfNormCostLayer(LayerBase):
         self.config.softmax_selfnorm_alpha = softmax_selfnorm_alpha
 
 
+@config_layer('cross_entropy_over_beam')
+class CrossEntropyOverBeamLayer(LayerBase):
+    def __init__(self, name, inputs, **xargs):
+        config_assert(len(inputs) % 3 == 0, "Error input number.")
+        super(CrossEntropyOverBeamLayer, self).__init__(
+            name, 'cross_entropy_over_beam', 0, inputs, **xargs)
+        input_num = len(inputs) / 3
+        for i in range(input_num):
+            input_layer = self.get_input_layer(i * 3)
+            config_assert(input_layer.size == 1, (
+                "Inputs for this layer are made up of "
+                "several triples, in which the first one is scores over "
+                "all candidate paths, whose size should be equal to 1."))
+
+
 @config_layer('fc')
 class FCLayer(LayerBase):
     layer_type = 'fc'
@@ -2448,6 +2463,7 @@ def define_cost(class_name, cost_type):
 
 
 define_cost('MultiClassCrossEntropy', 'multi-class-cross-entropy')
+define_cost('CrossEntropyOverBeamCostLayer', 'cross_entropy_over_beam')
 define_cost('RankingCost', 'rank-cost')
 define_cost('AucValidation', 'auc-validation')
 define_cost('PnpairValidation', 'pnpair-validation')

python/paddle/trainer_config_helpers/layers.py

@@ -11,7 +11,6 @@
 # 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.
-
 import functools
 import collections
 import inspect
@@ -106,6 +105,8 @@ __all__ = [
     'nce_layer',
     'cross_entropy_with_selfnorm',
     'cross_entropy',
+    'BeamInput',
+    'cross_entropy_over_beam',
     'multi_binary_label_cross_entropy',
     'sum_cost',
     'rank_cost',
@@ -229,6 +230,7 @@ class LayerType(object):
     HUBER_CLASSIFICATION = 'huber_classification'
     CROSS_ENTROPY = 'multi-class-cross-entropy'
     CROSS_ENTROPY_WITH_SELFNORM = 'multi_class_cross_entropy_with_selfnorm'
+    CROSS_ENTROPY_OVER_BEAM = 'cross_entropy_over_beam'
     SOFT_BIN_CLASS_CROSS_ENTROPY = 'soft_binary_class_cross_entropy'
     MULTI_BIN_LABEL_CROSS_ENTROPY = 'multi_binary_label_cross_entropy'
     SUM_COST = 'sum_cost'
@@ -4079,8 +4081,12 @@ def __cost_input__(input, label, weight=None):
     """
     inputs and parents for cost layers.
     """
-    ipts = [Input(input.name), Input(label.name)]
-    parents = [input, label]
+    if isinstance(input, LayerOutput):
+        input = [input]
+    if isinstance(label, LayerOutput):
+        label = [label]
+    ipts = [Input(ipt.name) for ipt in (input + label)]
+    parents = [ipt for ipt in (input + label)]
     if weight is not None:
         assert weight.size == 1
         ipts.append(Input(weight.name))
@@ -5067,17 +5073,6 @@ def warp_ctc_layer(input,
     building process, PaddlePaddle will clone the source codes, build and
     install it to :code:`third_party/install/warpctc` directory.
 
-    To use warp_ctc layer, you need to specify the path of :code:`libwarpctc.so`,
-    using following methods:
-
-    1. Set it in :code:`paddle.init` (python api) or :code:`paddle_init` (c api),
-    such as :code:`paddle.init(use_gpu=True,
-    warpctc_dir=your_paddle_source_dir/third_party/install/warpctc/lib)`.
-
-    2. Set environment variable LD_LIBRARY_PATH on Linux or DYLD_LIBRARY_PATH
-    on Mac OS. For instance, :code:`export
-    LD_LIBRARY_PATH=your_paddle_source_dir/third_party/install/warpctc/lib:$LD_LIBRARY_PATH`.
-
     More details of CTC can be found by referring to `Connectionist Temporal
     Classification: Labelling Unsegmented Sequence Data with Recurrent
     Neural Networks <http://machinelearning.wustl.edu/mlpapers/paper_files/
@@ -5785,10 +5780,10 @@ def multi_binary_label_cross_entropy(input,
 
     if input.activation is None or \
             not isinstance(input.activation, SigmoidActivation):
-        logger.log(
-            logging.WARN,
-            "%s is not recommend for multi_binary_label_cross_entropy's activation, "
-            "maybe the sigmoid is better" % repr(input.activation))
+        logger.log(logging.WARN,
+                   ("%s is not a recommended activation for "
+                    "multi_binary_label_cross_entropy, sigmoid is better") %
+                   repr(input.activation))
 
     Layer(
         name=name,
@@ -5803,6 +5798,113 @@ def multi_binary_label_cross_entropy(input,
         size=1)
 
 
+class BeamInput(object):
+    """
+    Define the input for cross_entropy_over_beam layer.
+
+    A beam is made up of a triple: the first one is scores over all
+    candidates; the second one is indices of top k selected candidates; the
+    third one is the index of ground truth, which is also always called
+    gold.
+    """
+
+    def __init__(self, candidate_scores, selected_candidates, gold):
+        assert isinstance(candidate_scores, LayerOutput)
+        self.candidate_scores = candidate_scores
+        assert candidate_scores.size == 1
+
+        assert isinstance(selected_candidates, LayerOutput)
+        self.selected_candidates = selected_candidates
+
+        assert isinstance(gold, LayerOutput)
+        self.gold = gold
+
+
+@wrap_name_default()
+@layer_support()
+def cross_entropy_over_beam(input, name=None):
+    """
+    This layer is used in learning to search models, which is to solve complex
+    joint prediction problems based on learning to search through a
+    problem-defined search space.
+
+    Specifically, the learning to search process for this layer begins with
+    searching a target sequence from a nested sequence. In the first search
+    step, top beam size sequences with highest scores, indices of these top k
+    sequences in the original nested sequence, and the ground truth (also
+    called gold) altogether (a triple) make up of the first beam.
+
+    Then, several special positions, for example, start and end positions
+    that define meaningful segments are searched. In these searches, top k
+    positions with highest scores are selected, and then sequence, starting
+    from the selected starts till ends of the sequences (or a fixed position)
+    are taken to search next.
+
+    We call the possible top k results returned in one search the beam. This
+    search process can be repeated for pre-defined turns and leads to several
+    beam expansions.
+
+    Finally, the layer cross_entropy_over_beam takes all the beam expansions
+    which contain several candidate targets found along the multi-step search.
+    cross_entropy_over_beam calculates cross entropy over the expanded beams
+    which all the candidates in the beam as the normalized factor.
+
+    Note that, if gold falls off the beam at search step t, then the cost is
+    calculated over the beam at step t.
+
+    This cost layer always works together with kmax_sequence_score_layer,
+    sub_nested_seq_layer, and sequence_slice_layer to trim the input to form a
+    sub-search space.
+
+
+    The example usage is:
+
+    .. code-block:: python
+
+       cost = cross_entropy_over_beam(input=[
+           BeamInput(
+               candidate_scores=beam1_candidates,
+               selected_candidates=beam1_topk,
+               gold=gold1),
+           BeamInput(
+               candidate_scores=beam2_candidates,
+               selected_candidates=beam2_topk,
+               gold=gold2),
+       ])
+
+
+    :param input: input beams for this layer.
+    :type input: BeamInput
+    :param name: input beams for this layer.
+    :type name: basestring
+    :return: LayerOutput object.
+    :rtype: LayerOutput
+    """
+
+    if isinstance(input, BeamInput):
+        input = [input]
+    else:
+        assert isinstance(input, list), (
+            'input for cross_entropy_over_beam shold be a python list '
+            'of BeamInput object.')
+        for ipt in input:
+            assert isinstance(ipt, BeamInput), (
+                'input for cross_entropy_over_beam '
+                'should be a BeamInput object.')
+
+    ipts = []
+    parents = []
+    for beam in input:
+        parents += [beam.candidate_scores, beam.selected_candidates, beam.gold]
+        ipts += [
+            beam.candidate_scores.name, beam.selected_candidates.name,
+            beam.gold.name
+        ]
+
+    Layer(name=name, type=LayerType.CROSS_ENTROPY_OVER_BEAM, inputs=ipts)
+    return LayerOutput(name, LayerType.CROSS_ENTROPY, parents=parents, size=1)
+
+
 @wrap_name_default()
 @layer_support()
 def smooth_l1_cost(input, label, name=None, coeff=1.0, layer_attr=None):

python/paddle/trainer_config_helpers/tests/configs/file_list.sh

@@ -9,6 +9,6 @@ test_seq_concat_reshape test_pad test_smooth_l1 test_multiplex_layer
 test_prelu_layer test_row_conv test_detection_output_layer test_multibox_loss_layer
 test_recursive_topology test_gated_unit_layer test_clip_layer test_row_l2_norm_layer
 test_kmax_seq_socre_layer test_seq_select_layers test_scale_shift_layer
-test_seq_slice_layer test_conv3d_layer test_deconv3d_layer)
+test_seq_slice_layer test_cross_entropy_over_beam  test_conv3d_layer test_deconv3d_layer)
 
 export whole_configs=(test_split_datasource)

python/paddle/trainer_config_helpers/tests/configs/protostr/test_cross_entropy_over_beam.protostr

+type: "nn"
+layers {
+  name: "sentence_states"
+  type: "data"
+  size: 32
+  active_type: ""
+}
+layers {
+  name: "sentence_scores"
+  type: "data"
+  size: 1
+  active_type: ""
+}
+layers {
+  name: "__kmax_sequence_score_layer_0__"
+  type: "kmax_seq_score"
+  active_type: ""
+  inputs {
+    input_layer_name: "sentence_scores"
+  }
+  beam_size: 5
+}
+layers {
+  name: "__sub_nested_seq_layer_0__"
+  type: "sub_nested_seq"
+  size: 32
+  active_type: ""
+  inputs {
+    input_layer_name: "sentence_states"
+  }
+  inputs {
+    input_layer_name: "__kmax_sequence_score_layer_0__"
+  }
+}
+layers {
+  name: "__fc_layer_0__"
+  type: "fc"
+  size: 1
+  active_type: ""
+  inputs {
+    input_layer_name: "__sub_nested_seq_layer_0__"
+    input_parameter_name: "___fc_layer_0__.w0"
+  }
+  bias_parameter_name: "___fc_layer_0__.wbias"
+}
+layers {
+  name: "__kmax_sequence_score_layer_1__"
+  type: "kmax_seq_score"
+  active_type: ""
+  inputs {
+    input_layer_name: "sentence_scores"
+  }
+  beam_size: 5
+}
+layers {
+  name: "__seq_slice_layer_0__"
+  type: "seq_slice"
+  size: 32
+  active_type: ""
+  inputs {
+    input_layer_name: "__sub_nested_seq_layer_0__"
+  }
+  inputs {
+    input_layer_name: "__kmax_sequence_score_layer_1__"
+  }
+  select_first: true
+}
+layers {
+  name: "__fc_layer_1__"
+  type: "fc"
+  size: 1
+  active_type: ""
+  inputs {
+    input_layer_name: "__seq_slice_layer_0__"
+    input_parameter_name: "___fc_layer_1__.w0"
+  }
+  bias_parameter_name: "___fc_layer_1__.wbias"
+}
+layers {
+  name: "__kmax_sequence_score_layer_2__"
+  type: "kmax_seq_score"
+  active_type: ""
+  inputs {
+    input_layer_name: "__fc_layer_1__"
+  }
+  beam_size: 5
+}
+layers {
+  name: "sentences_ids"
+  type: "data"
+  size: 1
+  active_type: ""
+}
+layers {
+  name: "start_ids"
+  type: "data"
+  size: 1
+  active_type: ""
+}
+layers {
+  name: "end_ids"
+  type: "data"
+  size: 1
+  active_type: ""
+}
+layers {
+  name: "__cross_entropy_over_beam_0__"
+  type: "cross_entropy_over_beam"
+  active_type: ""
+  inputs {
+    input_layer_name: "sentence_scores"
+  }
+  inputs {
+    input_layer_name: "__kmax_sequence_score_layer_0__"
+  }
+  inputs {
+    input_layer_name: "sentences_ids"
+  }
+  inputs {
+    input_layer_name: "__fc_layer_0__"
+  }
+  inputs {
+    input_layer_name: "__kmax_sequence_score_layer_1__"
+  }
+  inputs {
+    input_layer_name: "start_ids"
+  }
+  inputs {
+    input_layer_name: "__fc_layer_1__"
+  }
+  inputs {
+    input_layer_name: "__kmax_sequence_score_layer_2__"
+  }
+  inputs {
+    input_layer_name: "end_ids"
+  }
+}
+parameters {
+  name: "___fc_layer_0__.w0"
+  size: 32
+  initial_mean: 0.0
+  initial_std: 0.176776695297
+  dims: 32
+  dims: 1
+  initial_strategy: 0
+  initial_smart: true
+}
+parameters {
+  name: "___fc_layer_0__.wbias"
+  size: 1
+  initial_mean: 0.0
+  initial_std: 0.0
+  dims: 1
+  dims: 1
+  initial_strategy: 0
+  initial_smart: false
+}
+parameters {
+  name: "___fc_layer_1__.w0"
+  size: 32
+  initial_mean: 0.0
+  initial_std: 0.176776695297
+  dims: 32
+  dims: 1
+  initial_strategy: 0
+  initial_smart: true
+}
+parameters {
+  name: "___fc_layer_1__.wbias"
+  size: 1
+  initial_mean: 0.0
+  initial_std: 0.0
+  dims: 1
+  dims: 1
+  initial_strategy: 0
+  initial_smart: false
+}
+input_layer_names: "sentence_scores"
+input_layer_names: "sentences_ids"
+input_layer_names: "sentence_states"
+input_layer_names: "start_ids"
+input_layer_names: "end_ids"
+output_layer_names: "__cross_entropy_over_beam_0__"
+sub_models {
+  name: "root"
+  layer_names: "sentence_states"
+  layer_names: "sentence_scores"
+  layer_names: "__kmax_sequence_score_layer_0__"
+  layer_names: "__sub_nested_seq_layer_0__"
+  layer_names: "__fc_layer_0__"
+  layer_names: "__kmax_sequence_score_layer_1__"
+  layer_names: "__seq_slice_layer_0__"
+  layer_names: "__fc_layer_1__"
+  layer_names: "__kmax_sequence_score_layer_2__"
+  layer_names: "sentences_ids"
+  layer_names: "start_ids"
+  layer_names: "end_ids"
+  layer_names: "__cross_entropy_over_beam_0__"
+  input_layer_names: "sentence_scores"
+  input_layer_names: "sentences_ids"
+  input_layer_names: "sentence_states"
+  input_layer_names: "start_ids"
+  input_layer_names: "end_ids"
+  output_layer_names: "__cross_entropy_over_beam_0__"
+  is_recurrent_layer_group: false
+}
+

python/paddle/trainer_config_helpers/tests/configs/test_cross_entropy_over_beam.py

+#!/usr/bin/env python
+#coding=utf-8
+
+from paddle.trainer_config_helpers import *
+beam_size = 5
+
+# the first beam expansion.
+sentence_states = data_layer(name="sentence_states", size=32)
+sentence_scores = data_layer(name="sentence_scores", size=1)
+topk_sentence_ids = kmax_sequence_score_layer(
+    input=sentence_scores, beam_size=beam_size)
+
+# the second beam expansion.
+topk_sen = sub_nested_seq_layer(
+    input=sentence_states, selected_indices=topk_sentence_ids)
+start_pos_scores = fc_layer(input=topk_sen, size=1, act=LinearActivation())
+topk_start_pos_ids = kmax_sequence_score_layer(
+    input=sentence_scores, beam_size=beam_size)
+
+# the final beam expansion.
+topk_start_spans = seq_slice_layer(
+    input=topk_sen, starts=topk_start_pos_ids, ends=None)
+end_pos_scores = fc_layer(
+    input=topk_start_spans, size=1, act=LinearActivation())
+topk_end_pos_ids = kmax_sequence_score_layer(
+    input=end_pos_scores, beam_size=beam_size)
+
+# define the cost
+sentence_idx = data_layer(name="sentences_ids", size=1)
+start_idx = data_layer(name="start_ids", size=1)
+end_idx = data_layer(name="end_ids", size=1)
+cost = cross_entropy_over_beam(input=[
+    BeamInput(
+        candidate_scores=sentence_scores,
+        selected_candidates=topk_sentence_ids,
+        gold=sentence_idx), BeamInput(
+            candidate_scores=start_pos_scores,
+            selected_candidates=topk_start_pos_ids,
+            gold=start_idx), BeamInput(
+                candidate_scores=end_pos_scores,
+                selected_candidates=topk_end_pos_ids,
+                gold=end_idx)
+])
+
+outputs(cost)

python/paddle/v2/inference.py

@@ -70,7 +70,7 @@ class Inference(object):
                 item = [each_result[each_field] for each_field in field]
                 yield item
 
-    def infer(self, input, field='value', **kwargs):
+    def infer(self, input, field='value', flatten_result=True, **kwargs):
         """
         Infer a data by model.
         :param input: input data batch. Should be python iterable object.
@@ -83,7 +83,10 @@ class Inference(object):
                 retv = [[] for i in xrange(len(result))]
             for i, item in enumerate(result):
                 retv[i].append(item)
-        retv = [numpy.concatenate(out) for out in retv]
+
+        if flatten_result:
+            retv = [numpy.concatenate(out) for out in retv]
+
         if len(retv) == 1:
             return retv[0]
         else: