!5010 [MS][LITE][Develop]enable batchnorm fp16

Merge pull request !5010 from sunsuodong/batch_norm_fp16

mindspore/lite/src/runtime/kernel/arm/fp16/batchnorm_fp16.cc

@@ -23,44 +23,75 @@ using mindspore::lite::KernelRegistrar;
 using mindspore::schema::PrimitiveType_BatchNorm;
 
 namespace mindspore::kernel {
-int BatchnormFp16CPUKernel::DoExecute(int task_id) {
-  auto param = reinterpret_cast<BatchNormParameter *>(op_parameter_);
-
-  if (in_tensors_.at(0)->data_type() == kNumberTypeFloat32) {
-    auto input = in_tensors_.at(0);
-    auto mean = in_tensors_.at(1);
-    auto variance = in_tensors_.at(2);
-    auto output = out_tensors_.at(0);
+int BatchnormFp16CPUKernel::InitConstTensor() {
+  isFloat32Tensor_ = in_tensors_.at(0)->data_type() == kNumberTypeFloat32;
+  if (isFloat32Tensor_) {
+    auto mean_fp32 = in_tensors_.at(1);
+    auto variance_fp32 = in_tensors_.at(2);
+    mean_ = malloc(mean_fp32->ElementsNum() * sizeof(float16_t));
+    variance_ = malloc(variance_fp32->ElementsNum() * sizeof(float16_t));
+    if (mean_ == nullptr || variance_ == nullptr) {
+      FreeMeanAndVariance();
+      return RET_ERROR;
+    }
+    Float32ToFloat16(reinterpret_cast<float *>(mean_fp32->Data()),
+                     reinterpret_cast<float16_t *>(mean_), mean_fp32->ElementsNum());
+    Float32ToFloat16(reinterpret_cast<float *>(variance_fp32->Data()),
+                     reinterpret_cast<float16_t *>(variance_), variance_fp32->ElementsNum());
+  } else {
+    BatchnormCPUKernel::InitConstTensor();
+  }
+  return RET_OK;
+}
 
-    auto input_fp16 = context_->allocator->Malloc(input->ElementsNum() * sizeof(float16_t));
-    auto mean_fp16 = context_->allocator->Malloc(mean->ElementsNum() * sizeof(float16_t));
-    auto variance_fp16 = context_->allocator->Malloc(variance->ElementsNum() * sizeof(float16_t));
-    auto output_fp16 = context_->allocator->Malloc(output->ElementsNum() * sizeof(float16_t));
-    if (input_fp16 == nullptr || mean_fp16 == nullptr || variance_fp16 == nullptr || output_fp16 == nullptr) {
-      context_->allocator->Free(input_fp16);
-      context_->allocator->Free(mean_fp16);
-      context_->allocator->Free(variance_fp16);
-      context_->allocator->Free(output_fp16);
+int BatchnormFp16CPUKernel::Run() {
+  auto ret = Prepare();
+  if (ret != RET_OK) {
+    MS_LOG(ERROR) << "Prepare fail! Ret error code: " << ret;
+    return ret;
+  }
+  auto input_fp32 = in_tensors_.at(0);
+  auto output_fp32 = out_tensors_.at(0);
+  if (isFloat32Tensor_) {
+    input_ = context_->allocator->Malloc(input_fp32->ElementsNum() * sizeof(float16_t));
+    output_ = context_->allocator->Malloc(output_fp32->ElementsNum() * sizeof(float16_t));
+    if (input_ == nullptr || output_ == nullptr) {
+      FreeInputAndOutput();
+      return RET_ERROR;
     }
-    Float32ToFloat16(reinterpret_cast<float *>(input->Data()),
-                     reinterpret_cast<float16_t *>(input_fp16), input->ElementsNum());
-    Float32ToFloat16(reinterpret_cast<float *>(mean->Data()),
-                     reinterpret_cast<float16_t *>(mean_fp16), mean->ElementsNum());
-    Float32ToFloat16(reinterpret_cast<float *>(variance->Data()),
-                     reinterpret_cast<float16_t *>(variance_fp16), variance->ElementsNum());
+    Float32ToFloat16(reinterpret_cast<float *>(input_fp32->Data()),
+                  reinterpret_cast<float16_t *>(input_), input_fp32->ElementsNum());
+  } else {
+    input_ = in_tensors_.at(0)->Data();
+    output_ = out_tensors_.at(0)->Data();
+  }
+  ret = LiteBackendParallelLaunch(BatchNormRun, this, op_parameter_->thread_num_);
+  if (ret != RET_OK) {
+    MS_LOG(ERROR) << "BatchnormRun error error_code[" << ret << "]";
+  }
+  if (isFloat32Tensor_) {
+    Float16ToFloat32(reinterpret_cast<float16_t *>(output_), reinterpret_cast<float *>(output_fp32->Data()),
+                     output_fp32->ElementsNum());
+    FreeInputAndOutput();
+  }
+  return ret;
+}
 
-    BatchNormFp16(input_fp16, mean_fp16, variance_fp16, param, task_id, output_fp16);
+int BatchnormFp16CPUKernel::DoExecute(int task_id) {
+  auto param = reinterpret_cast<BatchNormParameter *>(op_parameter_);
+  BatchNormFp16(input_, mean_, variance_, param, task_id, output_);
+  return mindspore::lite::RET_OK;
+}
 
-    Float16ToFloat32(reinterpret_cast<float16_t *>(output_fp16), reinterpret_cast<float *>(output),
-                     output->ElementsNum());
-    context_->allocator->Free(input_fp16);
-    context_->allocator->Free(mean_fp16);
-    context_->allocator->Free(variance_fp16);
-    context_->allocator->Free(output_fp16);
-    return mindspore::lite::RET_OK;
+void BatchnormFp16CPUKernel::FreeInputAndOutput() {
+  if (input_ != nullptr) {
+    context_->allocator->Free(input_);
+    input_ = nullptr;
+  }
+  if (output_ != nullptr) {
+    context_->allocator->Free(output_);
+    output_ = nullptr;
   }
-  BatchNormFp16(in_tensors_.at(0)->Data(), mean_, variance_, param, task_id, out_tensors_.at(0)->Data());
-  return mindspore::lite::RET_OK;
 }
 
 kernel::LiteKernel *CpuBatchnormFp16KernelCreator(const std::vector<lite::tensor::Tensor *> &inputs,
@@ -83,5 +114,5 @@ kernel::LiteKernel *CpuBatchnormFp16KernelCreator(const std::vector<lite::tensor
   return kernel;
 }
 
-// REG_KERNEL(kCPU, kNumberTypeFloat16, PrimitiveType_BatchNorm, CpuBatchnormFp16KernelCreator)
+REG_KERNEL(kCPU, kNumberTypeFloat16, PrimitiveType_BatchNorm, CpuBatchnormFp16KernelCreator)
 }  // namespace mindspore::kernel

mindspore/lite/src/runtime/kernel/arm/fp16/batchnorm_fp16.h

@@ -29,7 +29,15 @@ class BatchnormFp16CPUKernel : public BatchnormCPUKernel {
       : BatchnormCPUKernel(parameter, inputs, outputs, ctx, primitive) {}
   virtual ~BatchnormFp16CPUKernel() {}
 
-  virtual int DoExecute(int task_id);
+  int Run() override;
+  int InitConstTensor() override;
+  int DoExecute(int task_id) override;
+
+ private:
+  void FreeInputAndOutput();
+  bool isFloat32Tensor_ = false;
+  void *input_ = nullptr;
+  void *output_ = nullptr;
 };
 }  // namespace mindspore::kernel