!5776 support caffe exp op

Merge pull request !5776 from 陶云浩/master

mindspore/lite/nnacl/fp32/arithmetic_self.c

@@ -34,14 +34,6 @@ int ElementCos(float *input, float *output, int element_size) {
   return NNACL_OK;
 }
 
-// exp:
-int ElementExp(float *input, float *output, int element_size) {
-  for (int i = 0; i < element_size; i++) {
-    output[i] = expf(input[i]);
-  }
-  return NNACL_OK;
-}
-
 // log:
 int ElementLog(float *input, float *output, int element_size) {
   for (int i = 0; i < element_size; i++) {

mindspore/lite/nnacl/fp32/arithmetic_self.h

@@ -30,8 +30,6 @@ int ElementAbs(float *input, float *output, int element_size);
 
 int ElementCos(float *input, float *output, int element_size);
 
-int ElementExp(float *input, float *output, int element_size);
-
 int ElementLog(float *input, float *output, int element_size);
 
 int ElementSquare(float *input, float *output, int element_size);

mindspore/lite/nnacl/fp32/exp.c

@@ -0,0 +1,37 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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 "nnacl/fp32/exp.h"
+#include <math.h>
+#include "nnacl/errorcode.h"
+
+int Exp(float *input_data, float *output_data, ExpParameter *parameter, int task_id) {
+  if (parameter->scale_ == 1) {
+    for (size_t i = task_id; i < parameter->element_num_; i += parameter->thread_num_) {
+      output_data[i] = expf(input_data[i]);
+    }
+  } else {
+    for (size_t i = task_id; i < parameter->element_num_; i += parameter->thread_num_) {
+      output_data[i] = expf(input_data[i] * parameter->in_scale_);
+    }
+  }
+  if (parameter->out_scale_ != 1) {
+    for (size_t i = task_id; i < parameter->element_num_; i += parameter->thread_num_) {
+      output_data[i] = output_data[i] * parameter->out_scale_;
+    }
+  }
+  return NNACL_OK;
+}

mindspore/lite/nnacl/fp32/exp.h

@@ -0,0 +1,41 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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.
+ */
+
+#ifndef MINDSPORE_LITE_NNACL_FP32_EXP_H_
+#define MINDSPORE_LITE_NNACL_FP32_EXP_H_
+
+#include "nnacl/op_base.h"
+
+typedef struct ExpParameter {
+  OpParameter op_parameter_;
+  int thread_num_;
+  float base_;
+  float scale_;
+  float shift_;
+  float in_scale_;
+  float out_scale_;
+  int element_num_;
+} ExpParameter;
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+int Exp(float *input_data, float *output_data, ExpParameter *parameter, int task_id);
+#ifdef __cplusplus
+}
+#endif
+
+#endif  // MINDSPORE_LITE_NNACL_FP32_EXP_H_

mindspore/lite/schema/ops.fbs

@@ -478,9 +478,9 @@ table Neg {
 }
 
 table Exp {
-    base : float;
-    scale : float;
-    shift : float;
+    base : float = -1.0;
+    scale : float = 1.0;
+    shift : float = 0.0;
 }
 
 table Cos {

mindspore/lite/src/ops/exp.cc

@@ -18,16 +18,35 @@
 
 namespace mindspore {
 namespace lite {
-#ifndef PRIMITIVE_WRITEABLE
+#ifdef PRIMITIVE_WRITEABLE
+void Exp::SetBase(float base) { this->primitive_->value.AsExp()->base = base; }
+void Exp::SetScale(float scale) { this->primitive_->value.AsExp()->scale = scale; }
+void Exp::SetShift(float shift) { this->primitive_->value.AsExp()->shift = shift; }
+
+float Exp::GetBase() const { return this->primitive_->value.AsExp()->base; }
+float Exp::GetScale() const { return this->primitive_->value.AsExp()->scale; }
+float Exp::GetShift() const { return this->primitive_->value.AsExp()->shift; }
+#else
+
 int Exp::UnPackToFlatBuilder(const schema::Primitive *primitive, flatbuffers::FlatBufferBuilder *fbb) {
   MS_ASSERT(nullptr != primitive);
   MS_ASSERT(nullptr != fbb);
-  auto val_offset = schema::CreateExp(*fbb);
+  auto attr = primitive->value_as_Exp();
+  if (attr == nullptr) {
+    MS_LOG(ERROR) << "value_as_Exp return nullptr";
+    return RET_ERROR;
+  }
+
+  auto val_offset = schema::CreateExp(*fbb, attr->base(), attr->scale(), attr->shift());
   auto prim_offset = schema::CreatePrimitive(*fbb, schema::PrimitiveType_Exp, val_offset.o);
   fbb->Finish(prim_offset);
   return RET_OK;
 }
 
+float Exp::GetBase() const { return this->primitive_->value_as_Exp()->base(); }
+float Exp::GetScale() const { return this->primitive_->value_as_Exp()->scale(); }
+float Exp::GetShift() const { return this->primitive_->value_as_Exp()->shift(); }
+
 #endif
 }  // namespace lite
 }  // namespace mindspore

mindspore/lite/src/ops/exp.h

@@ -21,21 +21,27 @@
 #include <set>
 #include <cmath>
 #include "ir/dtype/type_id.h"
-#include "src/ops/arithmetic_self.h"
+#include "src/ops/primitive_c.h"
 
 namespace mindspore {
 namespace lite {
-class Exp : public ArithmeticSelf {
+class Exp : public PrimitiveC {
  public:
 #ifdef PRIMITIVE_WRITEABLE
-  MS_DECLARE_PARENT(Exp, ArithmeticSelf);
+  MS_DECLARE_PARENT(Exp, PrimitiveC);
   Exp() = default;
-  explicit Exp(schema::PrimitiveT *primitive) : ArithmeticSelf(primitive) {}
+  explicit Exp(schema::PrimitiveT *primitive) : PrimitiveC(primitive) {}
+  void SetBase(float base);
+  void SetShift(float shift);
+  void SetScale(float scale);
 #else
   Exp() = default;
 
   int UnPackToFlatBuilder(const schema::Primitive *primitive, flatbuffers::FlatBufferBuilder *fbb) override;
 #endif
+  float GetBase() const;
+  float GetShift() const;
+  float GetScale() const;
 };
 }  // namespace lite
 }  // namespace mindspore

mindspore/lite/src/ops/floor_div.cc

@@ -23,7 +23,7 @@ namespace lite {
 int FloorDiv::UnPackToFlatBuilder(const schema::Primitive *primitive, flatbuffers::FlatBufferBuilder *fbb) {
   MS_ASSERT(nullptr != primitive);
   MS_ASSERT(nullptr != fbb);
-  auto val_offset = schema::CreateFloor(*fbb);
+  auto val_offset = schema::CreateFloorDiv(*fbb);
   auto prim_offset = schema::CreatePrimitive(*fbb, schema::PrimitiveType_FloorDiv, val_offset.o);
   fbb->Finish(prim_offset);
   return RET_OK;

mindspore/lite/src/populate_parameter.cc

@@ -173,6 +173,7 @@
 #include "nnacl/sparse_to_dense.h"
 #include "nnacl/l2_norm_parameter.h"
 #include "nnacl/detection_post_process_parameter.h"
+#include "nnacl/fp32/exp.h"
 
 namespace mindspore::kernel {
 
@@ -1519,8 +1520,7 @@ OpParameter *PopulateEluParameter(const mindspore::lite::PrimitiveC *primitive)
   return reinterpret_cast<OpParameter *>(elu_parameter);
 }
 
-OpParameter *PopulateL2NormParameter(
-    const mindspore::lite::PrimitiveC *primitive) {
+OpParameter *PopulateL2NormParameter(const mindspore::lite::PrimitiveC *primitive) {
   L2NormParameter *l2_norm_parameter = reinterpret_cast<L2NormParameter *>(malloc(sizeof(L2NormParameter)));
   if (l2_norm_parameter == nullptr) {
     MS_LOG(ERROR) << "malloc L2NormParameter failed.";
@@ -1568,6 +1568,26 @@ OpParameter *PopulateDetectionPostProcessParameter(const mindspore::lite::Primit
   return reinterpret_cast<OpParameter *>(detection_post_process_parameter);
 }
 
+OpParameter *PopulateExpParameter(const mindspore::lite::PrimitiveC *primitive) {
+  ExpParameter *exp_parameter = reinterpret_cast<ExpParameter *>(malloc(sizeof(ExpParameter)));
+  if (exp_parameter == nullptr) {
+    MS_LOG(ERROR) << "malloc ExpParameter failed.";
+    return nullptr;
+  }
+  memset(exp_parameter, 0, sizeof(ExpParameter));
+  exp_parameter->op_parameter_.type_ = primitive->Type();
+  auto param = reinterpret_cast<mindspore::lite::Exp *>(const_cast<mindspore::lite::PrimitiveC *>(primitive));
+  exp_parameter->base_ = param->GetBase();
+  exp_parameter->scale_ = param->GetScale();
+  exp_parameter->shift_ = param->GetShift();
+  if (exp_parameter->base_ != -1 && exp_parameter->base_ <= 0) {
+    MS_LOG(ERROR) << "Exp base must be strictly positive, got " << exp_parameter->base_;
+    free(exp_parameter);
+    return nullptr;
+  }
+  return reinterpret_cast<OpParameter *>(exp_parameter);
+}
+
 PopulateParameterRegistry::PopulateParameterRegistry() {
   populate_parameter_funcs_[schema::PrimitiveType_SparseToDense] = PopulateSparseToDenseParameter;
   populate_parameter_funcs_[schema::PrimitiveType_SoftMax] = PopulateSoftmaxParameter;
@@ -1610,7 +1630,7 @@ PopulateParameterRegistry::PopulateParameterRegistry() {
   populate_parameter_funcs_[schema::PrimitiveType_Abs] = PopulateArithmeticSelf;
   populate_parameter_funcs_[schema::PrimitiveType_Cos] = PopulateArithmeticSelf;
   populate_parameter_funcs_[schema::PrimitiveType_Sin] = PopulateArithmeticSelf;
-  populate_parameter_funcs_[schema::PrimitiveType_Exp] = PopulateArithmeticSelf;
+  populate_parameter_funcs_[schema::PrimitiveType_Exp] = PopulateExpParameter;
   populate_parameter_funcs_[schema::PrimitiveType_Log] = PopulateArithmeticSelf;
   populate_parameter_funcs_[schema::PrimitiveType_Square] = PopulateArithmeticSelf;
   populate_parameter_funcs_[schema::PrimitiveType_Sqrt] = PopulateArithmeticSelf;

mindspore/lite/src/runtime/kernel/arm/fp32/arithmetic_self.cc

@@ -169,7 +169,6 @@ kernel::LiteKernel *CpuArithmeticSelfFp32KernelCreator(const std::vector<lite::t
 
 REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Abs, CpuArithmeticSelfFp32KernelCreator)
 REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Cos, CpuArithmeticSelfFp32KernelCreator)
-REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Exp, CpuArithmeticSelfFp32KernelCreator)
 REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Log, CpuArithmeticSelfFp32KernelCreator)
 REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Square, CpuArithmeticSelfFp32KernelCreator)
 REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Sqrt, CpuArithmeticSelfFp32KernelCreator)

mindspore/lite/src/runtime/kernel/arm/fp32/arithmetic_self.h

@@ -28,7 +28,6 @@ using mindspore::lite::Context;
 using mindspore::schema::PrimitiveType_Abs;
 using mindspore::schema::PrimitiveType_Ceil;
 using mindspore::schema::PrimitiveType_Cos;
-using mindspore::schema::PrimitiveType_Exp;
 using mindspore::schema::PrimitiveType_Floor;
 using mindspore::schema::PrimitiveType_Log;
 using mindspore::schema::PrimitiveType_LogicalNot;
@@ -55,9 +54,6 @@ class ArithmeticSelfCPUKernel : public LiteKernel {
       case PrimitiveType_Cos:
         arithmeticSelf_run_ = ElementCos;
         break;
-      case PrimitiveType_Exp:
-        arithmeticSelf_run_ = ElementExp;
-        break;
       case PrimitiveType_Log:
         arithmeticSelf_run_ = ElementLog;
         break;

mindspore/lite/src/runtime/kernel/arm/fp32/exp.cc

@@ -0,0 +1,115 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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 "src/runtime/kernel/arm/fp32/exp.h"
+#include <math.h>
+#include "include/errorcode.h"
+#include "src/kernel_registry.h"
+#include "src/runtime/runtime_api.h"
+
+using mindspore::lite::KernelRegistrar;
+using mindspore::lite::RET_ERROR;
+using mindspore::lite::RET_OK;
+using mindspore::schema::PrimitiveType_Exp;
+
+namespace mindspore::kernel {
+int ExpCPUKernel::Init() {
+  exp_parameter_ = reinterpret_cast<ExpParameter *>(op_parameter_);
+  exp_parameter_->thread_num_ = thread_count_;
+
+  if (!InferShapeDone()) {
+    return RET_OK;
+  }
+
+  return ReSize();
+}
+
+int ExpCPUKernel::ReSize() {
+  exp_parameter_->thread_num_ = thread_count_;
+  float log_ = (exp_parameter_->base_ == -1) ? 1 : logf(exp_parameter_->base_);
+  exp_parameter_->in_scale_ = exp_parameter_->scale_ * log_;
+  if (exp_parameter_->shift_ == 0) {
+    exp_parameter_->out_scale_ = 1;
+  } else {
+    if (log_ == 1) {
+      exp_parameter_->out_scale_ = expf(exp_parameter_->shift_);
+    } else {
+      exp_parameter_->out_scale_ = powf(exp_parameter_->base_, exp_parameter_->shift_);
+    }
+  }
+  return RET_OK;
+}
+
+int ExpCPUKernel::DoExcute(int task_id) {
+  Exp(input_addr_, output_addr_, exp_parameter_, task_id);
+  return RET_OK;
+}
+
+int ExpRun(void *cdata, int task_id) {
+  auto ExpData = reinterpret_cast<ExpCPUKernel *>(cdata);
+  auto ret = ExpData->DoExcute(task_id);
+  if (ret != RET_OK) {
+    MS_LOG(ERROR) << "ExpRun error task_id[" << task_id << "] error_code[" << ret << "]";
+    return RET_ERROR;
+  }
+  return RET_OK;
+}
+
+int ExpCPUKernel::Run() {
+  auto prepare_ret = Prepare();
+  if (prepare_ret != RET_OK) {
+    MS_LOG(ERROR) << "Prepare fail!ret: " << prepare_ret;
+    return prepare_ret;
+  }
+  input_addr_ = reinterpret_cast<float *>(in_tensors_.front()->Data());
+  output_addr_ = reinterpret_cast<float *>(out_tensors_.front()->Data());
+  exp_parameter_->element_num_ = in_tensors_.front()->ElementsNum();
+
+  auto ret = ParallelLaunch(THREAD_POOL_DEFAULT, ExpRun, this, exp_parameter_->thread_num_);
+  if (ret != RET_OK) {
+    MS_LOG(ERROR) << "Exp error: error_code[" << ret << "]";
+    return RET_ERROR;
+  }
+  return RET_OK;
+}
+
+kernel::LiteKernel *CpuExpFp32KernelCreator(const std::vector<lite::tensor::Tensor *> &inputs,
+                                            const std::vector<lite::tensor::Tensor *> &outputs, OpParameter *parameter,
+                                            const lite::Context *ctx, const KernelKey &desc,
+                                            const mindspore::lite::PrimitiveC *primitive) {
+  if (parameter == nullptr || ctx == nullptr) {
+    MS_LOG(ERROR) << "parameter or ctx is nullptr";
+    return nullptr;
+  }
+  MS_ASSERT(desc.type == PrimitiveType_Exp);
+  auto *kernel = new (std::nothrow) ExpCPUKernel(parameter, inputs, outputs, ctx, primitive);
+  if (kernel == nullptr) {
+    MS_LOG(ERROR) << "Create Kernel failed, name: " << parameter->name_;
+    return nullptr;
+  }
+
+  auto ret = kernel->Init();
+  if (ret != RET_OK) {
+    MS_LOG(ERROR) << "Init Kernel failed, name: " << parameter->name_
+                  << ", type: " << schema::EnumNamePrimitiveType(static_cast<schema::PrimitiveType>(parameter->type_));
+    delete kernel;
+    return nullptr;
+  }
+  return kernel;
+}
+
+REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_Exp, CpuExpFp32KernelCreator)
+}  // namespace mindspore::kernel

mindspore/lite/src/runtime/kernel/arm/fp32/exp.h

@@ -0,0 +1,49 @@
+/**
+ * Copyright 2020 Huawei Technologies Co., Ltd
+ *
+ * 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.
+ */
+
+#ifndef MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_FP32_EXP_H_
+#define MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_FP32_EXP_H_
+
+#include <vector>
+#include "src/lite_kernel.h"
+#include "nnacl/fp32/exp.h"
+
+namespace mindspore::kernel {
+class ExpCPUKernel : public LiteKernel {
+ public:
+  explicit ExpCPUKernel(OpParameter *parameter, const std::vector<lite::tensor::Tensor *> &inputs,
+                        const std::vector<lite::tensor::Tensor *> &outputs, const lite::Context *ctx,
+                        const mindspore::lite::PrimitiveC *primitive)
+      : LiteKernel(parameter, inputs, outputs, ctx, primitive), ctx_(ctx), thread_count_(ctx->thread_num_) {}
+  ~ExpCPUKernel() override{};
+
+  int Init() override;
+  int ReSize() override;
+  int Run() override;
+  int DoExcute(int task_id);
+
+ protected:
+  const lite::Context *ctx_;
+  int thread_count_;
+  ExpParameter *exp_parameter_;
+
+ private:
+  float *input_addr_;
+  float *output_addr_;
+};
+}  // namespace mindspore::kernel
+
+#endif  // MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_FP32_EXP_H_