Merge pull request #1390 from hjchen2/ocr_attention

Aggregate activation operations, add log and lod_reset op

src/common/types.cpp

@@ -73,6 +73,8 @@ const char *G_OP_TYPE_SHAPE = "shape";
 const char *G_OP_TYPE_SUM = "sum";
 const char *G_OP_TYPE_TOP_K = "top_k";
 const char *G_OP_TYPE_CAST = "cast";
+const char *G_OP_TYPE_LOG = "log";
+const char *G_OP_TYPE_LOD_RESET = "lod_reset";
 
 const char *G_OP_TYPE_QUANTIZE = "quantize";
 const char *G_OP_TYPE_DEQUANTIZE = "dequantize";
@@ -171,5 +173,7 @@ std::unordered_map<
         {G_OP_TYPE_SEQUENCE_EXPAND, {{"X", "Y"}, {"Out"}}},
         {G_OP_TYPE_SEQUENCE_POOL, {{"X"}, {"Out"}}},
         {G_OP_TYPE_SEQUENCE_SOFTMAX, {{"X"}, {"Out"}}},
-        {G_OP_TYPE_NORM, {{"X"}, {"Out", "Norm"}}}};
+        {G_OP_TYPE_NORM, {{"X"}, {"Out", "Norm"}}},
+        {G_OP_TYPE_LOG, {{"X"}, {"Out"}}},
+        {G_OP_TYPE_LOD_RESET, {{"X", "Y"}, {"Out"}}}};
 }  // namespace paddle_mobile

src/common/types.h

@@ -100,6 +100,7 @@ enum ActivationType {
   LEAKY_RELU = 4,
   TANH = 5,
   SIGMOID = 6,
+  LOG = 7,
 };
 
 enum PoolingType {
@@ -155,6 +156,8 @@ extern const char *G_OP_TYPE_PRELU;
 extern const char *G_OP_TYPE_SUM;
 extern const char *G_OP_TYPE_TOP_K;
 extern const char *G_OP_TYPE_CAST;
+extern const char *G_OP_TYPE_LOG;
+extern const char *G_OP_TYPE_LOD_RESET;
 
 extern const char *G_OP_TYPE_QUANTIZE;
 extern const char *G_OP_TYPE_DEQUANTIZE;

src/framework/load_ops.h

@@ -270,3 +270,6 @@ LOAD_OP1(sequence_expand, CPU);
 #ifdef SEQUENCE_POOL_OP
 LOAD_OP1(sequence_pool, CPU);
 #endif
+#ifdef LOG_OP
+LOAD_OP1(log, CPU);
+#endif

src/framework/operator.h

@@ -36,13 +36,12 @@ limitations under the License. */
 #include "framework/cl/cl_helper.h"
 #include "framework/cl/cl_scope.h"
 #endif
+
 namespace paddle_mobile {
 namespace framework {
-using std::string;
-using std::vector;
 
 template <typename T>
-static T *GetVarValue(const string &key, const VariableNameMap &var_map,
+static T *GetVarValue(const std::string &key, const VariableNameMap &var_map,
                       const Scope &scope) {
   auto var_vec = var_map.at(key);
   if (!var_vec.empty()) {
@@ -56,44 +55,29 @@ static T *GetVarValue(const string &key, const VariableNameMap &var_map,
 template <typename Dtype>
 class OperatorBase {
  public:
-  /*
-   *  @b op 基类的实例化方法, op 获取到了 输入、参数以及提前分配好的输出 tensor
-   * */
   OperatorBase(const std::string &type, const VariableNameMap &inputs,
                const VariableNameMap &outputs, const AttributeMap &attrs,
                std::shared_ptr<Scope> scope);
   virtual ~OperatorBase() {}
-  void Run();
-  std::vector<string> GetOutKeys() const;
-  std::vector<string> GetInputKeys() const;
-  virtual void RunImpl() = 0;
 
   virtual void Init() = 0;
-  /*
-   * @b op 运算所需的输入, 如上一层的输出结果、卷积核
-   * */
+  virtual void InferShape() const = 0;
+  virtual void Run();
+  virtual void RunImpl() = 0;
+
+  std::vector<std::string> GetOutKeys() const;
+  std::vector<std::string> GetInputKeys() const;
+
   const VariableNameMap &Inputs() const { return inputs_; }
-  /*
-   * @b op 的输出, 内存会提前被分配好, 运算结果会被存到分配好的内存内
-   * */
   const VariableNameMap &Outputs() const { return outputs_; }
-  /*
-   * @b op 类型
-   * */
   const std::string &Type() const { return type_; }
-  /*
-   * @b op 运算所需要用到的参数: 如 conv 运算所需要用到的 stride
-   * */
   const AttributeMap &Attrs() const { return attrs_; }
+
   void ClearVariables(const std::vector<std::string> &var_names) const {
     if (this->scope_) {
       this->scope_->EraseVars(var_names);
     }
   }
-  /*
-   * @b 根据输入形状和参数计算出输出形状
-   * */
-  virtual void InferShape() const = 0;
 
  protected:
   std::shared_ptr<Scope> scope_;
@@ -106,9 +90,6 @@ class OperatorBase {
   void CheckAllInputOutputSet() const;
 };
 
-/*
- * @b 这个类为所有带有运算的 op 的父类, 这个 op 继承与 OperatorBase
- * */
 template <typename Dtype, typename ParamType, typename KernelType>
 class OperatorWithKernel : public OperatorBase<Dtype> {
  public:
@@ -136,9 +117,6 @@ class OperatorWithKernel : public OperatorBase<Dtype> {
   ParamType param_;
 };
 
-/*
- * @b 所有kernel的父类
- * */
 template <typename Dtype, typename P>
 class OpKernelBase {
  public:
@@ -150,11 +128,6 @@ class OpKernelBase {
   }
 #endif
 
-    /*
-     * @b 所有kernel 需实现 Compute 方法
-     * @p para 这个参数为 kernel 运算时所需要用到参数组成的一个结构体,
-     *    所有结构体存在与: paddle-mobile/src/operators/op_param.h
-     * */
 #ifdef PADDLE_McOBILE_MALI_GPU
   OpKernelBase() { acl_op_ = nullptr; }
   void *GetAclOp() const { return acl_op_; }
@@ -177,13 +150,6 @@ class OpKernelBase {
 #endif
 };
 
-#define DEFINE_OP_CONSTRUCTOR(cls, parent_cls)                                 \
-  cls(const std::string &type, const ::paddle_mobile::VariableNameMap &inputs, \
-      const ::paddle_mobile::VariableNameMap &outputs,                         \
-      const ::paddle_mobile::framework::AttributeMap &attrs,                   \
-      std::shared_ptr<::paddle_mobile::framework::Scope> scope)                \
-      : parent_cls<Dtype, T>(type, inputs, outputs, attrs, scope) {}
-
 class FusionOpMatcher {
  public:
   FusionOpMatcher() {}
@@ -202,12 +168,44 @@ class FusionOpMatcher {
 
   virtual std::vector<std::pair<int, std::string>> NeedCheck() { return {}; }
 
-  //  virtual  bool Fusion();
  protected:
   Node node_;
   std::string type_;
   std::shared_ptr<OpDesc> new_opdesc_;
 };
 
+#define DECLARE_OPERATOR(OpName, OpParam, OpKernel)                          \
+  template <typename DeviceType, typename T>                                 \
+  class OpName##Op : public framework::OperatorWithKernel<                   \
+                         DeviceType, OpParam<DeviceType>,                    \
+                         operators::OpKernel<DeviceType, T>> {               \
+   public:                                                                   \
+    OpName##Op(const std::string &type, const VariableNameMap &inputs,       \
+               const VariableNameMap &outputs,                               \
+               const framework::AttributeMap &attrs,                         \
+               std::shared_ptr<framework::Scope> scope)                      \
+        : framework::OperatorWithKernel<DeviceType, OpParam<DeviceType>,     \
+                                        operators::OpKernel<DeviceType, T>>( \
+              type, inputs, outputs, attrs, scope) {}                        \
+                                                                             \
+    void InferShape() const override;                                        \
+  };
+
+#define DECLARE_KERNEL(OpName, OpParam)                                   \
+  template <typename DeviceType, typename T>                              \
+  class OpName##Kernel                                                    \
+      : public framework::OpKernelBase<DeviceType, OpParam<DeviceType>> { \
+   public:                                                                \
+    bool Init(OpParam<DeviceType> *param);                                \
+    void Compute(const OpParam<DeviceType> &param);                       \
+  };
+
+#define DEFINE_OP_CONSTRUCTOR(cls, parent_cls)                                 \
+  cls(const std::string &type, const ::paddle_mobile::VariableNameMap &inputs, \
+      const ::paddle_mobile::VariableNameMap &outputs,                         \
+      const ::paddle_mobile::framework::AttributeMap &attrs,                   \
+      std::shared_ptr<::paddle_mobile::framework::Scope> scope)                \
+      : parent_cls<Dtype, T>(type, inputs, outputs, attrs, scope) {}
+
 }  // namespace framework
 }  // namespace paddle_mobile

src/operators/relu_op.cpp → src/operators/activation_op.cpp

@@ -12,29 +12,40 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
-#ifdef RELU_OP
-
-#include "operators/relu_op.h"
+#include "operators/activation_op.h"
 
 namespace paddle_mobile {
 namespace operators {
 
-template <typename Dtype, typename T>
-void ReluOp<Dtype, T>::InferShape() const {
-  auto input_dims = this->param_.InputX()->dims();
-  this->param_.Out()->Resize(input_dims);
-}
+#define DEFINE_INFERSHAPE(OpName)                           \
+  template <typename Dtype, typename T>                     \
+  void OpName##Op<Dtype, T>::InferShape() const {           \
+    const auto &input_dims = this->param_.InputX()->dims(); \
+    this->param_.Out()->Resize(input_dims);                 \
+  }
+
+#ifdef RELU_OP
+DEFINE_INFERSHAPE(Relu);
+DEFINE_INFERSHAPE(Relu6);
+#endif  // RELU_OP
 
-template <typename Dtype, typename T>
-void Relu6Op<Dtype, T>::InferShape() const {
-  auto input_dims = this->param_.InputX()->dims();
-  this->param_.Out()->Resize(input_dims);
-}
+#ifdef SIGMOID_OP
+DEFINE_INFERSHAPE(Sigmoid);
+#endif  // SIGMOID_OP
+
+#ifdef TANH_OP
+DEFINE_INFERSHAPE(Tanh);
+#endif  // TANH_OP
+
+#ifdef LOG_OP
+DEFINE_INFERSHAPE(Log);
+#endif  // LOG_OP
 
 }  // namespace operators
 }  // namespace paddle_mobile
 
 namespace ops = paddle_mobile::operators;
+#ifdef RELU_OP
 #ifdef PADDLE_MOBILE_CPU
 REGISTER_OPERATOR_CPU(relu, ops::ReluOp);
 REGISTER_OPERATOR_CPU(relu6, ops::Relu6Op);
@@ -47,5 +58,23 @@ REGISTER_OPERATOR_MALI_GPU(relu, ops::ReluOp);
 #ifdef PADDLE_MOBILE_CL
 REGISTER_OPERATOR_CL(relu, ops::ReluOp);
 #endif
-
 #endif  // RELU_OP
+
+#ifdef SIGMOID_OP
+#ifdef PADDLE_MOBILE_CPU
+REGISTER_OPERATOR_CPU(sigmoid, ops::SigmoidOp);
+#endif
+#endif  // SIGMOID_OP
+
+#ifdef TANH_OP
+#ifdef PADDLE_MOBILE_CPU
+REGISTER_OPERATOR_CPU(tanh, ops::TanhOp);
+#endif
+#ifdef PADDLE_MOBILE_FPGA
+REGISTER_OPERATOR_FPGA(tanh, ops::TanhOp);
+#endif
+#endif  // TANH_OP
+
+#ifdef LOG_OP
+REGISTER_OPERATOR_CPU(log, ops::LogOp);
+#endif  // LOG_OP

src/operators/tanh_op.h → src/operators/activation_op.h

@@ -12,8 +12,6 @@ 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. */
 
-#ifdef TANH_OP
-
 #pragma once
 
 #include <string>
@@ -24,21 +22,22 @@ limitations under the License. */
 namespace paddle_mobile {
 namespace operators {
 
-template <typename DeviceType, typename T>
-class TanhOp : public framework::OperatorWithKernel<
-                   DeviceType, TanhParam<DeviceType>,
-                   operators::TanhKernel<DeviceType, T>> {
- public:
-  TanhOp(const std::string &type, const VariableNameMap &inputs,
-         const VariableNameMap &outputs, const framework::AttributeMap &attrs,
-         std::shared_ptr<framework::Scope> scope)
-      : framework::OperatorWithKernel<DeviceType, TanhParam<DeviceType>,
-                                      operators::TanhKernel<DeviceType, T>>(
-            type, inputs, outputs, attrs, scope) {}
-  void InferShape() const override;
-};
+#ifdef RELU_OP
+DECLARE_OPERATOR(Relu, ReluParam, ReluKernel);
+DECLARE_OPERATOR(Relu6, ReluParam, Relu6Kernel);
+#endif
 
-}  // namespace operators
-}  // namespace paddle_mobile
+#ifdef SIGMOID_OP
+DECLARE_OPERATOR(Sigmoid, SigmoidParam, SigmoidKernel);
+#endif
 
+#ifdef TANH_OP
+DECLARE_OPERATOR(Tanh, TanhParam, TanhKernel);
+#endif
+
+#ifdef LOG_OP
+DECLARE_OPERATOR(Log, ReluParam, LogKernel);
 #endif
+
+}  // namespace operators
+}  // namespace paddle_mobile

src/operators/kernel/activation_kernel.h

@@ -20,26 +20,21 @@ limitations under the License. */
 namespace paddle_mobile {
 namespace operators {
 
-#define DECLARE_KERNEL(KernelClass, KernelParam)                              \
-  template <typename DeviceType, typename T>                                  \
-  class KernelClass                                                           \
-      : public framework::OpKernelBase<DeviceType, KernelParam<DeviceType>> { \
-   public:                                                                    \
-    bool Init(KernelParam<DeviceType> *param);                                \
-    void Compute(const KernelParam<DeviceType> &param);                       \
-  };
-
 #ifdef RELU_OP
-DECLARE_KERNEL(ReluKernel, ReluParam);
-DECLARE_KERNEL(Relu6Kernel, ReluParam);
+DECLARE_KERNEL(Relu, ReluParam);
+DECLARE_KERNEL(Relu6, ReluParam);
 #endif
 
 #ifdef SIGMOID_OP
-DECLARE_KERNEL(SigmoidKernel, SigmoidParam);
+DECLARE_KERNEL(Sigmoid, SigmoidParam);
 #endif
 
 #ifdef TANH_OP
-DECLARE_KERNEL(TanhKernel, TanhParam);
+DECLARE_KERNEL(Tanh, TanhParam);
+#endif
+
+#ifdef LOG_OP
+DECLARE_KERNEL(Log, ReluParam);
 #endif
 
 }  // namespace operators

src/operators/kernel/arm/activation_kernel.cpp

@@ -105,7 +105,7 @@ void SigmoidKernel<CPU, float>::Compute(const SigmoidParam<CPU> &param) {
 
 #ifdef TANH_OP
 template <>
-void TanhKernel<CPU, float>::Init(TanhParam<CPU> *param) {
+bool TanhKernel<CPU, float>::Init(TanhParam<CPU> *param) {
   return true;
 }
 
@@ -117,5 +117,19 @@ void TanhKernel<CPU, float>::Compute(const TanhParam<CPU> &param) {
 }
 #endif
 
+#ifdef LOG_OP
+template <>
+bool LogKernel<CPU, float>::Init(ReluParam<CPU> *param) {
+  return true;
+}
+
+template <>
+void LogKernel<CPU, float>::Compute(const ReluParam<CPU> &param) {
+  const Tensor *input = param.InputX();
+  Tensor *output = param.Out();
+  ActivationCompute<float, LOG>()(input, output);
+}
+#endif
+
 }  // namespace operators
 }  // namespace paddle_mobile

src/operators/sigmoid_op.h → src/operators/kernel/arm/lod_reset_kernel.cpp

@@ -12,34 +12,51 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
-#ifdef SIGMOID_OP
+#ifdef LOD_RESET_OP
 
-#pragma once
-
-#include <string>
-#include "framework/operator.h"
-#include "operators/kernel/activation_kernel.h"
-#include "operators/op_param.h"
+#include "operators/kernel/kernels.h"
 
 namespace paddle_mobile {
 namespace operators {
 
-template <typename DeviceType, typename T>
-class SigmoidOp : public framework::OperatorWithKernel<
-                      DeviceType, SigmoidParam<DeviceType>,
-                      operators::SigmoidKernel<DeviceType, T>> {
- public:
-  SigmoidOp(const std::string &type, const VariableNameMap &inputs,
-            const VariableNameMap &outputs,
-            const framework::AttributeMap &attrs,
-            std::shared_ptr<framework::Scope> scope)
-      : framework::OperatorWithKernel<DeviceType, SigmoidParam<DeviceType>,
-                                      operators::SigmoidKernel<DeviceType, T>>(
-            type, inputs, outputs, attrs, scope) {}
-  void InferShape() const override;
-};
+template <>
+bool LodResetKernel<CPU, float>::Init(LodResetParam<CPU> *param) {
+  return true;
+}
+
+template <>
+void LodResetKernel<CPU, float>::Compute(const LodResetParam<CPU> &param) {
+  const auto *input = param.input_x_;
+  const auto *lod_t = param.input_y_;
+  auto *output = param.output_;
+
+  output->ShareDataWith(*input);
+
+  std::vector<int> level0;
+  if (lod_t) {
+    if (lod_t->lod().size() > 0) {
+      output->set_lod(lod_t->lod());
+      return;  // early return, since lod already set
+    } else {
+      auto *lod = lod_t->data<int>();
+      level0 = std::vector<int>(lod, lod + lod_t->numel());
+    }
+  } else {
+    level0 = param.target_lod_;
+  }
+
+  // cast level0 to size_t
+  std::vector<size_t> ulevel0(level0.size(), 0);
+  for (int i = 0; i < level0.size(); ++i) {
+    ulevel0[i] = level0[i];
+  }
+
+  framework::LoD target_lod;
+  target_lod.push_back(std::move(ulevel0));
+  output->set_lod(target_lod);
+}
 
 }  // namespace operators
 }  // namespace paddle_mobile
 
-#endif
+#endif  // LOD_RESET_OP

src/operators/kernel/arm/sequence_pool_kernel.cpp

@@ -66,8 +66,9 @@ void SequencePoolImpl(const framework::LoDTensor &input,
       memcpy(out_ptr, in_ptr, width * sizeof(float));
       in_ptr += width;
       int remain_h = height - 1;
+      int remain_w_start = 0;
 #ifdef __ARM_NEON__
-      int remain_w_start = width & 0xfffc;
+      remain_w_start = width & 0xfffc;
 #endif  // __ARM_NEON__
       for (int h = 0; h < remain_h; ++h) {
 #ifdef __ARM_NEON__
@@ -124,9 +125,10 @@ void SequencePoolImpl<SUM, float>(const framework::LoDTensor &input,
       memcpy(out_ptr, in_ptr, width * sizeof(float));
       in_ptr += width;
       int remain_h = height - 1;
+      int remain_w_start = 0;
 #ifdef __ARM_NEON__
       int loop_w = width >> 2;
-      int remain_w_start = width & 0xfffc;
+      remain_w_start = width & 0xfffc;
 #endif  // __ARM_NEON__
       for (int h = 0; h < remain_h; ++h) {
 #ifdef __ARM_NEON__

src/operators/kernel/dequant_bn_kernel.h

@@ -20,37 +20,28 @@ limitations under the License. */
 namespace paddle_mobile {
 namespace operators {
 
-#define DECLARE_KERNEL(KernelClass, KernelParam)                              \
-  template <typename DeviceType, typename T>                                  \
-  class KernelClass                                                           \
-      : public framework::OpKernelBase<DeviceType, KernelParam<DeviceType>> { \
-   public:                                                                    \
-    bool Init(KernelParam<DeviceType> *param);                                \
-    void Compute(const KernelParam<DeviceType> &param);                       \
-  };
-
 #ifdef FUSION_DEQUANT_BN_OP
-DECLARE_KERNEL(FusionDequantBNKernel, FusionDequantBNParam);
+DECLARE_KERNEL(FusionDequantBN, FusionDequantBNParam);
 #endif
 
 #ifdef FUSION_DEQUANT_BN_RELU_OP
-DECLARE_KERNEL(FusionDequantBNReluKernel, FusionDequantBNParam);
+DECLARE_KERNEL(FusionDequantBNRelu, FusionDequantBNParam);
 #endif
 
 #ifdef FUSION_DEQUANT_ADD_BN_OP
-DECLARE_KERNEL(FusionDequantAddBNKernel, FusionDequantAddBNParam);
+DECLARE_KERNEL(FusionDequantAddBN, FusionDequantAddBNParam);
 #endif
 
 #ifdef FUSION_DEQUANT_ADD_BN_RELU_OP
-DECLARE_KERNEL(FusionDequantAddBNReluKernel, FusionDequantAddBNParam);
+DECLARE_KERNEL(FusionDequantAddBNRelu, FusionDequantAddBNParam);
 #endif
 
 #ifdef FUSION_DEQUANT_ADD_BN_QUANT_OP
-DECLARE_KERNEL(FusionDequantAddBNQuantKernel, FusionDequantAddBNQuantParam);
+DECLARE_KERNEL(FusionDequantAddBNQuant, FusionDequantAddBNQuantParam);
 #endif
 
 #ifdef FUSION_DEQUANT_ADD_BN_RELU_QUANT_OP
-DECLARE_KERNEL(FusionDequantAddBNReluQuantKernel, FusionDequantAddBNQuantParam);
+DECLARE_KERNEL(FusionDequantAddBNReluQuant, FusionDequantAddBNQuantParam);
 #endif
 
 }  // namespace operators

src/operators/kernel/kernels.h

@@ -20,22 +20,17 @@ limitations under the License. */
 namespace paddle_mobile {
 namespace operators {
 
-#define DECLARE_KERNEL(KernelClass, KernelParam)                              \
-  template <typename DeviceType, typename T>                                  \
-  class KernelClass                                                           \
-      : public framework::OpKernelBase<DeviceType, KernelParam<DeviceType>> { \
-   public:                                                                    \
-    bool Init(KernelParam<DeviceType> *param);                                \
-    void Compute(const KernelParam<DeviceType> &param);                       \
-  };
-
 #ifdef TOP_K_OP
-DECLARE_KERNEL(TopKKernel, TopKParam)
+DECLARE_KERNEL(TopK, TopKParam)
 #endif  // TOP_K_OP
 
 #ifdef CAST_OP
-DECLARE_KERNEL(CastKernel, CastParam)
+DECLARE_KERNEL(Cast, CastParam)
 #endif  // CAST_OP
 
+#ifdef LOD_RESET_OP
+DECLARE_KERNEL(LodReset, LodResetParam)
+#endif  // LOD_RESET_OP
+
 }  // namespace operators
 }  // namespace paddle_mobile

src/operators/kernel/sequence_kernels.h

@@ -20,25 +20,16 @@ limitations under the License. */
 namespace paddle_mobile {
 namespace operators {
 
-#define DECLARE_KERNEL(KernelClass, KernelParam)                              \
-  template <typename DeviceType, typename T>                                  \
-  class KernelClass                                                           \
-      : public framework::OpKernelBase<DeviceType, KernelParam<DeviceType>> { \
-   public:                                                                    \
-    bool Init(KernelParam<DeviceType> *param);                                \
-    void Compute(const KernelParam<DeviceType> &param);                       \
-  };
-
 #ifdef SEQUENCE_EXPAND_OP
-DECLARE_KERNEL(SequenceExpandKernel, SequenceExpandParam);
+DECLARE_KERNEL(SequenceExpand, SequenceExpandParam);
 #endif  // SEQUENCE_EXPAND_OP
 
 #ifdef SEQUENCE_POOL_OP
-DECLARE_KERNEL(SequencePoolKernel, SequencePoolParam);
+DECLARE_KERNEL(SequencePool, SequencePoolParam);
 #endif  // SEQUENCE_POOL_OP
 
 #ifdef SEQUENCE_SOFTMAX_OP
-DECLARE_KERNEL(SequenceSoftmaxKernel, SoftmaxParam);
+DECLARE_KERNEL(SequenceSoftmax, SoftmaxParam);
 #endif  // SEQUENCE_SOFTMAX_OP
 
 }  // namespace operators

src/operators/sigmoid_op.cpp → src/operators/lod_reset_op.cpp

@@ -12,16 +12,17 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
-#ifdef SIGMOID_OP
+#ifdef LOD_RESET_OP
 
-#include "operators/sigmoid_op.h"
+#include "operators/lod_reset_op.h"
 
 namespace paddle_mobile {
 namespace operators {
 
-template <typename DeviceType, typename T>
-void SigmoidOp<DeviceType, T>::InferShape() const {
-  this->param_.Out()->Resize(this->param_.InputX()->dims());
+template <typename Dtype, typename T>
+void LodResetOp<Dtype, T>::InferShape() const {
+  const auto &input_dims = this->param_.input_->dims();
+  this->param_.output_->Resize(input_dims);
 }
 
 }  // namespace operators
@@ -29,7 +30,7 @@ void SigmoidOp<DeviceType, T>::InferShape() const {
 
 namespace ops = paddle_mobile::operators;
 #ifdef PADDLE_MOBILE_CPU
-REGISTER_OPERATOR_CPU(sigmoid, ops::SigmoidOp);
+REGISTER_OPERATOR_CPU(lod_reset, ops::LodResetOp);
 #endif
 
-#endif
+#endif  // LOD_RESET_OP

src/operators/tanh_op.cpp → src/operators/lod_reset_op.h

@@ -12,27 +12,21 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
-#ifdef TANH_OP
+#ifdef LOD_RESET_OP
 
-#include "operators/tanh_op.h"
+#pragma once
+
+#include <string>
+#include "framework/operator.h"
+#include "operators/kernel/kernels.h"
+#include "operators/op_param.h"
 
 namespace paddle_mobile {
 namespace operators {
 
-template <typename DeviceType, typename T>
-void TanhOp<DeviceType, T>::InferShape() const {
-  this->param_.Out()->Resize(this->param_.InputX()->dims());
-}
+DECLARE_OPERATOR(LodReset, LodResetParam, LodResetKernel);
 
 }  // namespace operators
 }  // namespace paddle_mobile
 
-namespace ops = paddle_mobile::operators;
-#ifdef PADDLE_MOBILE_CPU
-REGISTER_OPERATOR_CPU(tanh, ops::TanhOp);
-#endif
-#ifdef PADDLE_MOBILE_FPGA
-REGISTER_OPERATOR_FPGA(tanh, ops::TanhOp);
-#endif
-
-#endif
+#endif  // LOD_RESET_OP

src/operators/math/activation.h

@@ -86,6 +86,11 @@ inline float32x4_t vActiveq_f32<TANH>(const float32x4_t &x) {
   __out = vmulq_n_f32(__out, 2.f);
   return vsubq_f32(__out, __one);
 }
+
+template <>
+inline float32x4_t vActiveq_f32<LOG>(const float32x4_t &x) {
+  return log_ps(x);
+}
 #endif
 
 template <ActivationType Act = IDENTITY>
@@ -119,6 +124,11 @@ inline float Active<TANH>(const float &x) {
   return 2.f / (1.f + exp(-2.f * x)) - 1.f;
 }
 
+template <>
+inline float Active<LOG>(const float &x) {
+  return log(x);
+}
+
 }  // namespace math
 }  // namespace operators
 }  // namespace paddle_mobile

src/operators/op_param.h

@@ -2829,5 +2829,32 @@ class SequencePoolParam : public OpParam {
 };
 #endif  // SEQUENCE_EXPAND_OP
 
+#ifdef LOD_RESET_OP
+template <typename Dtype>
+class LodResetParam : public OpParam {
+  typedef typename DtypeTensorTrait<Dtype>::gtype GType;
+  typedef typename DtypeTensorTrait<Dtype>::rtype RType;
+
+ public:
+  LodResetParam(const VariableNameMap &inputs, const VariableNameMap &outputs,
+                const AttributeMap &attrs, const Scope &scope) {
+    input_x_ = InputXFrom<GType>(inputs, scope);
+    output_ = OutFrom<GType>(outputs, scope);
+    input_y_ = nullptr;
+    if (inputs.count("Y")) {
+      input_y_ = InputYFrom<GType>(inputs, scope);
+    } else {
+      target_lod_ = OpParam::GetAttr<vector<int>>("target_lod", attrs);
+    }
+  }
+
+ public:
+  GType *input_x_;
+  GType *input_y_;
+  GType *output_;
+  std::vector<int> target_lod_;
+};
+#endif  // LOD_RESET_OP
+
 }  // namespace operators
 }  // namespace paddle_mobile

src/operators/relu_op.h

-/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
-
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
-
-    http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License. */
-
-#ifdef RELU_OP
-
-#pragma once
-
-#include <string>
-
-#include "framework/operator.h"
-#include "operators/kernel/activation_kernel.h"
-#include "operators/op_param.h"
-
-namespace paddle_mobile {
-namespace operators {
-
-template <typename DeviceType, typename T>
-class ReluOp : public framework::OperatorWithKernel<
-                   DeviceType, ReluParam<DeviceType>,
-                   operators::ReluKernel<DeviceType, T>> {
- public:
-  ReluOp(const std::string &type, const VariableNameMap &inputs,
-         const VariableNameMap &outputs, const framework::AttributeMap &attrs,
-         std::shared_ptr<framework::Scope> scope)
-      : framework::OperatorWithKernel<DeviceType, ReluParam<DeviceType>,
-                                      operators::ReluKernel<DeviceType, T>>(
-            type, inputs, outputs, attrs, scope) {}
-
-  void InferShape() const override;
-};
-
-template <typename DeviceType, typename T>
-class Relu6Op : public framework::OperatorWithKernel<
-                    DeviceType, ReluParam<DeviceType>,
-                    operators::Relu6Kernel<DeviceType, T>> {
- public:
-  Relu6Op(const std::string &type, const VariableNameMap &inputs,
-          const VariableNameMap &outputs, const framework::AttributeMap &attrs,
-          std::shared_ptr<framework::Scope> scope)
-      : framework::OperatorWithKernel<DeviceType, ReluParam<DeviceType>,
-                                      operators::Relu6Kernel<DeviceType, T>>(
-            type, inputs, outputs, attrs, scope) {}
-
-  void InferShape() const override;
-};
-
-}  // namespace operators
-}  // namespace paddle_mobile
-
-#endif

test/CMakeLists.txt

@@ -238,6 +238,12 @@ if (NOT FOUND_MATCH)
     ADD_EXECUTABLE(test-relu6-op operators/test_relu6_op.cpp test_helper.h test_include.h)
     target_link_libraries(test-relu6-op paddle-mobile)
 
+    ADD_EXECUTABLE(test-tanh-op operators/test_tanh_op.cpp test_helper.h test_include.h)
+    target_link_libraries(test-tanh-op paddle-mobile)
+
+    ADD_EXECUTABLE(test-log-op operators/test_log_op.cpp test_helper.h test_include.h)
+    target_link_libraries(test-log-op paddle-mobile)
+
     ADD_EXECUTABLE(test-topk-op operators/test_topk_op.cpp test_helper.h test_include.h)
     target_link_libraries(test-topk-op paddle-mobile)
 

test/executor_for_test.h

@@ -20,12 +20,11 @@ limitations under the License. */
 #include "common/log.h"
 #include "framework/executor.h"
 #include "framework/op_registry.h"
+#include "operators/activation_op.h"
 #include "operators/conv_op.h"
 #include "operators/elementwise_add_op.h"
 #include "operators/pool_op.h"
-#include "operators/relu_op.h"
 #include "operators/reshape_op.h"
-#include "operators/sigmoid_op.h"
 #include "operators/softmax_op.h"
 #include "operators/transpose_op.h"
 

test/net/test_benchmark.cpp

@@ -43,7 +43,8 @@ int main(int argc, char* argv[]) {
     std::shared_ptr<paddle_mobile::framework::Tensor> output;
     std::vector<int64_t> dims{1, 3, 224, 224};
     if (feed_shape) {
-      sscanf(feed_shape, "%d,%d,%d,%d", &dims[0], &dims[1], &dims[2], &dims[3]);
+      sscanf(feed_shape, "%ld,%ld,%ld,%ld", &dims[0], &dims[1], &dims[2],
+             &dims[3]);
     }
     std::cout << "feed shape: [" << dims[0] << ", " << dims[1] << ", "
               << dims[2] << ", " << dims[3] << "]\n";

test/operators/test_log_op.cpp

+/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include <cmath>
+#include <iostream>
+#include "../test_include.h"
+#include "operators/activation_op.h"
+
+namespace paddle_mobile {
+
+void Log(const framework::Tensor *X, framework::Tensor *Y) {
+  const float *x = X->data<float>();
+  float *y = Y->mutable_data<float>();
+
+  for (int i = 0; i < X->numel(); ++i) {
+    y[i] = log(x[i]);
+  }
+}
+
+int TestLogOp(const std::vector<int> input_shape) {
+  framework::DDim dims = framework::make_ddim(input_shape);
+  VariableNameMap inputs;
+  VariableNameMap outputs;
+  auto scope = std::make_shared<framework::Scope>();
+  inputs["X"] = std::vector<std::string>({"input"});
+  outputs["Out"] = std::vector<std::string>({"output"});
+
+  auto input_var = scope.get()->Var("input");
+  auto input = input_var->template GetMutable<framework::LoDTensor>();
+  SetupTensor<float>(input, dims, 0.0001, 100.0);
+
+  auto output_var = scope.get()->Var("output");
+
+  framework::AttributeMap attrs;
+  auto *op =
+      new operators::LogOp<CPU, float>("log", inputs, outputs, attrs, scope);
+  op->InferShape();
+  op->Init();
+  op->Run();
+
+  auto output = output_var->template Get<framework::LoDTensor>();
+
+  framework::Tensor output_cmp;
+  float *output_cmp_data = output_cmp.mutable_data<float>(output->dims());
+  Log(input, &output_cmp);
+
+  const float *output_data = output->data<float>();
+  for (int i = 0; i < output->numel(); ++i) {
+    float gap = output_data[i] - output_cmp_data[i];
+    if (std::abs(gap / (output_data[i] + 1e-5)) > 1e-3) {
+      LOG(kLOG_INFO) << "output_data[" << i << "] = " << output_data[i]
+                     << ", output_cmp_data[" << i
+                     << "] = " << output_cmp_data[i];
+      delete op;
+      exit(1);
+    }
+  }
+  delete op;
+  return 0;
+}
+
+}  // namespace paddle_mobile
+
+int main() {
+  paddle_mobile::TestLogOp({1, 1, 2, 3});
+  paddle_mobile::TestLogOp({1, 3, 11, 22});
+  paddle_mobile::TestLogOp({1, 32, 112, 112});
+  std::cout << "test log op pass." << std::endl;
+  return 0;
+}

test/operators/test_relu6_op.cpp

@@ -15,7 +15,7 @@ limitations under the License. */
 #include <cmath>
 #include <iostream>
 #include "../test_include.h"
-#include "operators/relu_op.h"
+#include "operators/activation_op.h"
 
 namespace paddle_mobile {
 

test/operators/test_relu_op.cpp

@@ -15,7 +15,7 @@ limitations under the License. */
 #include <cmath>
 #include <iostream>
 #include "../test_include.h"
-#include "operators/relu_op.h"
+#include "operators/activation_op.h"
 
 namespace paddle_mobile {
 

test/operators/test_sigmoid_op.cpp

@@ -12,15 +12,70 @@ 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 "../test_helper.h"
+#include <cmath>
+#include <iostream>
+#include "../test_include.h"
+#include "operators/activation_op.h"
 
-int main() {
-  paddle_mobile::framework::Tensor input;
-  paddle_mobile::framework::Tensor output;
-  SetupTensor<float>(&input, {1, 4, 60, 60}, static_cast<float>(0),
-                     static_cast<float>(1));
+namespace paddle_mobile {
+
+void Sigmoid(const framework::Tensor *X, framework::Tensor *Y) {
+  const float *x = X->data<float>();
+  float *y = Y->mutable_data<float>();
+
+  for (int i = 0; i < X->numel(); ++i) {
+    y[i] = 1.f / (1.f + exp(-x[i]));
+  }
+}
+
+int TestSigmoidOp(const std::vector<int> input_shape) {
+  framework::DDim dims = framework::make_ddim(input_shape);
+  VariableNameMap inputs;
+  VariableNameMap outputs;
+  auto scope = std::make_shared<framework::Scope>();
+  inputs["X"] = std::vector<std::string>({"input"});
+  outputs["Out"] = std::vector<std::string>({"output"});
+
+  auto input_var = scope.get()->Var("input");
+  auto input = input_var->template GetMutable<framework::LoDTensor>();
+  SetupTensor<float>(input, dims, -100.0, 100.0);
+
+  auto output_var = scope.get()->Var("output");
+
+  framework::AttributeMap attrs;
+  auto *op = new operators::SigmoidOp<CPU, float>("sigmoid", inputs, outputs,
+                                                  attrs, scope);
+  op->InferShape();
+  op->Init();
+  op->Run();
 
-  auto out_ddim = paddle_mobile::framework::make_ddim({1, 4, 60, 60});
-  output.Resize(out_ddim);
+  auto output = output_var->template Get<framework::LoDTensor>();
+
+  framework::Tensor output_cmp;
+  float *output_cmp_data = output_cmp.mutable_data<float>(output->dims());
+  Sigmoid(input, &output_cmp);
+
+  const float *output_data = output->data<float>();
+  for (int i = 0; i < output->numel(); ++i) {
+    float gap = output_data[i] - output_cmp_data[i];
+    if (std::abs(gap / (output_data[i] + 1e-5)) > 1e-3) {
+      LOG(kLOG_INFO) << "output_data[" << i << "] = " << output_data[i]
+                     << ", output_cmp_data[" << i
+                     << "] = " << output_cmp_data[i];
+      delete op;
+      exit(1);
+    }
+  }
+  delete op;
+  return 0;
+}
+
+}  // namespace paddle_mobile
+
+int main() {
+  paddle_mobile::TestSigmoidOp({1, 1, 2, 3});
+  paddle_mobile::TestSigmoidOp({1, 3, 11, 22});
+  paddle_mobile::TestSigmoidOp({1, 32, 112, 112});
+  std::cout << "test sigmoid op pass." << std::endl;
   return 0;
 }

test/operators/test_tanh_op.cpp

+/* Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include <cmath>
+#include <iostream>
+#include "../test_include.h"
+#include "operators/activation_op.h"
+
+namespace paddle_mobile {
+
+void Tanh(const framework::Tensor *X, framework::Tensor *Y) {
+  const float *x = X->data<float>();
+  float *y = Y->mutable_data<float>();
+
+  for (int i = 0; i < X->numel(); ++i) {
+    y[i] = 2.f / (1.f + exp(-2.f * x[i])) - 1.f;
+  }
+}
+
+int TestTanhOp(const std::vector<int> input_shape) {
+  framework::DDim dims = framework::make_ddim(input_shape);
+  VariableNameMap inputs;
+  VariableNameMap outputs;
+  auto scope = std::make_shared<framework::Scope>();
+  inputs["X"] = std::vector<std::string>({"input"});
+  outputs["Out"] = std::vector<std::string>({"output"});
+
+  auto input_var = scope.get()->Var("input");
+  auto input = input_var->template GetMutable<framework::LoDTensor>();
+  SetupTensor<float>(input, dims, -100.0, 100.0);
+
+  auto output_var = scope.get()->Var("output");
+
+  framework::AttributeMap attrs;
+  auto *op =
+      new operators::TanhOp<CPU, float>("tanh", inputs, outputs, attrs, scope);
+  op->InferShape();
+  op->Init();
+  op->Run();
+
+  auto output = output_var->template Get<framework::LoDTensor>();
+
+  framework::Tensor output_cmp;
+  float *output_cmp_data = output_cmp.mutable_data<float>(output->dims());
+  Tanh(input, &output_cmp);
+
+  const float *output_data = output->data<float>();
+  for (int i = 0; i < output->numel(); ++i) {
+    float gap = output_data[i] - output_cmp_data[i];
+    if (std::abs(gap / (output_data[i] + 1e-5)) > 1e-3) {
+      LOG(kLOG_INFO) << "output_data[" << i << "] = " << output_data[i]
+                     << ", output_cmp_data[" << i
+                     << "] = " << output_cmp_data[i];
+      delete op;
+      exit(1);
+    }
+  }
+  delete op;
+  return 0;
+}
+
+}  // namespace paddle_mobile
+
+int main() {
+  paddle_mobile::TestTanhOp({1, 1, 2, 3});
+  paddle_mobile::TestTanhOp({1, 3, 11, 22});
+  paddle_mobile::TestTanhOp({1, 32, 112, 112});
+  std::cout << "test sigmoid op pass." << std::endl;
+  return 0;
+}

tools/op.cmake

@@ -276,6 +276,8 @@ if(NOT FOUND_MATCH)
   set(SEQUENCE_EXPAND_OP ON)
   set(SEQUENCE_POOL_OP ON)
   set(SEQUENCE_SOFTMAX_OP ON)
+  set(LOG_OP ON)
+  set(TANH_OP ON)
 endif()
 
   # option(BATCHNORM_OP "" ON)
@@ -512,6 +514,9 @@ endif()
 if (SEQUENCE_SOFTMAX_OP)
   add_definitions(-DSEQUENCE_SOFTMAX_OP)
 endif()
+if (LOG_OP)
+  add_definitions(-DLOG_OP)
+endif()
 
 if (TANH_OP)
   add_definitions(-DTANH_OP)