提交 6fc6f290 编写于 作者: M mindspore-ci-bot 提交者: Gitee

!3951 add Op_BatchNorm and testcase;

Merge pull request !3951 from songhonglei413/test
...@@ -53,7 +53,7 @@ union PrimitiveType { ...@@ -53,7 +53,7 @@ union PrimitiveType {
Activation, Activation,
Conv2D, Conv2D,
FusedBatchNorm, FusedBatchNorm,
CaffeBatchNorm, BatchNorm,
BiasAdd, BiasAdd,
Pooling, Pooling,
DepthwiseConv2D, DepthwiseConv2D,
......
...@@ -212,8 +212,8 @@ table Conv2DGradInput { ...@@ -212,8 +212,8 @@ table Conv2DGradInput {
spatial: int = 1; spatial: int = 1;
} }
table CaffeBatchNorm { table BatchNorm {
epsilon: float; // eg. epsilon=0.001 epsilon: float = 0.00001; // eg. epsilon=0.001
} }
table BiasGrad { table BiasGrad {
......
...@@ -37,7 +37,7 @@ constexpr const float POW_NUM = 0.5; ...@@ -37,7 +37,7 @@ constexpr const float POW_NUM = 0.5;
bool IsBatchNode(const BaseRef &n) { bool IsBatchNode(const BaseRef &n) {
if (utils::isa<CNodePtr>(n) || utils::isa<ValueNodePtr>(n)) { if (utils::isa<CNodePtr>(n) || utils::isa<ValueNodePtr>(n)) {
auto type = opt::GetCNodeType(n); auto type = opt::GetCNodeType(n);
return type == schema::PrimitiveType_CaffeBatchNorm || type == schema::PrimitiveType_FusedBatchNorm; return type == schema::PrimitiveType_BatchNorm || type == schema::PrimitiveType_FusedBatchNorm;
} }
return false; return false;
} }
...@@ -115,12 +115,12 @@ const void ConvBatchNormFusion::InitTransParam(const CNodePtr &bn_node, int kern ...@@ -115,12 +115,12 @@ const void ConvBatchNormFusion::InitTransParam(const CNodePtr &bn_node, int kern
AnfNodePtr bn_bias_node = nullptr; AnfNodePtr bn_bias_node = nullptr;
float eps = 0; float eps = 0;
auto primitiveT_value = GetValueNode<std::shared_ptr<lite::PrimitiveTValue>>(bn_node->input(0)); auto primitiveT_value = GetValueNode<std::shared_ptr<lite::PrimitiveTValue>>(bn_node->input(0));
if (GetCNodeType(bn_node) == schema::PrimitiveType_CaffeBatchNorm) { if (GetCNodeType(bn_node) == schema::PrimitiveType_BatchNorm) {
bn_mean_node = bn_node->input(kCaffeBNMeanIndex); bn_mean_node = bn_node->input(kCaffeBNMeanIndex);
bn_variance_node = bn_node->input(kCaffeBNVarIndex); bn_variance_node = bn_node->input(kCaffeBNVarIndex);
CheckIfNodeIsParam(bn_mean_node); CheckIfNodeIsParam(bn_mean_node);
CheckIfNodeIsParam(bn_variance_node); CheckIfNodeIsParam(bn_variance_node);
eps = primitiveT_value->GetPrimitiveT()->value.AsCaffeBatchNorm()->epsilon; eps = primitiveT_value->GetPrimitiveT()->value.AsBatchNorm()->epsilon;
} else if (GetCNodeType(bn_node) == schema::PrimitiveType_FusedBatchNorm) { } else if (GetCNodeType(bn_node) == schema::PrimitiveType_FusedBatchNorm) {
bn_scale_node = bn_node->input(kTFBNScaleIndex); bn_scale_node = bn_node->input(kTFBNScaleIndex);
bn_bias_node = bn_node->input(kTFBNBiasIndex); bn_bias_node = bn_node->input(kTFBNBiasIndex);
......
...@@ -90,8 +90,8 @@ lite::Primitive *ModelImpl::CopyPrimitive(const schema::Primitive *srcPrim) { ...@@ -90,8 +90,8 @@ lite::Primitive *ModelImpl::CopyPrimitive(const schema::Primitive *srcPrim) {
return new lite::DepthwiseConv2D(const_cast<schema::Primitive *>(srcPrim)); return new lite::DepthwiseConv2D(const_cast<schema::Primitive *>(srcPrim));
case schema::PrimitiveType_FusedBatchNorm: case schema::PrimitiveType_FusedBatchNorm:
return new lite::FusedBatchNorm(const_cast<schema::Primitive *>(srcPrim)); return new lite::FusedBatchNorm(const_cast<schema::Primitive *>(srcPrim));
case schema::PrimitiveType_CaffeBatchNorm: case schema::PrimitiveType_BatchNorm:
return new lite::CaffeBatchNorm(const_cast<schema::Primitive *>(srcPrim)); return new lite::BatchNorm(const_cast<schema::Primitive *>(srcPrim));
case schema::PrimitiveType_FullConnection: case schema::PrimitiveType_FullConnection:
return new lite::FullConnection(const_cast<schema::Primitive *>(srcPrim)); return new lite::FullConnection(const_cast<schema::Primitive *>(srcPrim));
case schema::PrimitiveType_Power: case schema::PrimitiveType_Power:
......
...@@ -39,8 +39,8 @@ Primitive *Primitive::CreatePrimitive(schema::Primitive *primitive) { ...@@ -39,8 +39,8 @@ Primitive *Primitive::CreatePrimitive(schema::Primitive *primitive) {
return new lite::DepthwiseConv2D(const_cast<schema::Primitive *>(primitive)); return new lite::DepthwiseConv2D(const_cast<schema::Primitive *>(primitive));
case schema::PrimitiveType_FusedBatchNorm: case schema::PrimitiveType_FusedBatchNorm:
return new lite::FusedBatchNorm(const_cast<schema::Primitive *>(primitive)); return new lite::FusedBatchNorm(const_cast<schema::Primitive *>(primitive));
case schema::PrimitiveType_CaffeBatchNorm: case schema::PrimitiveType_BatchNorm:
return new lite::CaffeBatchNorm(const_cast<schema::Primitive *>(primitive)); return new lite::BatchNorm(const_cast<schema::Primitive *>(primitive));
case schema::PrimitiveType_FullConnection: case schema::PrimitiveType_FullConnection:
return new lite::FullConnection(const_cast<schema::Primitive *>(primitive)); return new lite::FullConnection(const_cast<schema::Primitive *>(primitive));
case schema::PrimitiveType_Power: case schema::PrimitiveType_Power:
......
...@@ -90,10 +90,10 @@ class Pooling : public Primitive { ...@@ -90,10 +90,10 @@ class Pooling : public Primitive {
int pad_r_ = 0; int pad_r_ = 0;
}; };
class CaffeBatchNorm : public Primitive { class BatchNorm : public Primitive {
public: public:
explicit CaffeBatchNorm(schema::Primitive *primitive) : Primitive(primitive) {} explicit BatchNorm(schema::Primitive *primitive) : Primitive(primitive) {}
const schema::CaffeBatchNorm *GetAttribute() const { return this->primitive->value_as_CaffeBatchNorm(); } const schema::BatchNorm *GetAttribute() const { return this->primitive->value_as_BatchNorm(); }
}; };
class FusedBatchNorm : public Primitive { class FusedBatchNorm : public Primitive {
......
...@@ -39,6 +39,7 @@ ...@@ -39,6 +39,7 @@
#include "src/runtime/kernel/arm/opclib/fp32/activation.h" #include "src/runtime/kernel/arm/opclib/fp32/activation.h"
#include "src/runtime/kernel/arm/opclib/fp32/arithmetic.h" #include "src/runtime/kernel/arm/opclib/fp32/arithmetic.h"
#include "src/runtime/kernel/arm/opclib/fused_batchnorm.h" #include "src/runtime/kernel/arm/opclib/fused_batchnorm.h"
#include "src/runtime/kernel/arm/opclib/fp32/batchnorm.h"
#include "src/runtime/kernel/arm/opclib/power.h" #include "src/runtime/kernel/arm/opclib/power.h"
#include "src/runtime/kernel/arm/opclib/fp32/range.h" #include "src/runtime/kernel/arm/opclib/fp32/range.h"
#include "src/runtime/kernel/arm/opclib/fp32/local_response_norm.h" #include "src/runtime/kernel/arm/opclib/fp32/local_response_norm.h"
...@@ -70,6 +71,18 @@ ...@@ -70,6 +71,18 @@
#include "src/runtime/kernel/arm/opclib/fp32/lstm.h" #include "src/runtime/kernel/arm/opclib/fp32/lstm.h"
namespace mindspore::kernel { namespace mindspore::kernel {
OpParameter *PopulateBatchNorm(const lite::Primitive *primitive) {
BatchNormParameter *batch_norm_param = new (std::nothrow) BatchNormParameter();
if (batch_norm_param == nullptr) {
MS_LOG(ERROR) << "new BatchNormParameter failed.";
return nullptr;
}
batch_norm_param->op_parameter_.type_ = primitive->Type();
auto param = primitive->Value()->value_as_BatchNorm();
batch_norm_param->epsilon_ = param->epsilon();
return reinterpret_cast<OpParameter *>(batch_norm_param);
}
OpParameter *PopulateFillParameter(const lite::Primitive *primitive) { OpParameter *PopulateFillParameter(const lite::Primitive *primitive) {
auto param = primitive->Value()->value_as_Fill(); auto param = primitive->Value()->value_as_Fill();
FillParameter *fill_param = new (std::nothrow) FillParameter(); FillParameter *fill_param = new (std::nothrow) FillParameter();
...@@ -1190,6 +1203,7 @@ PopulateParameterRegistry::PopulateParameterRegistry() { ...@@ -1190,6 +1203,7 @@ PopulateParameterRegistry::PopulateParameterRegistry() {
populate_parameter_funcs_[schema::PrimitiveType_DeDepthwiseConv2D] = PopulateDeconvDwParameter; populate_parameter_funcs_[schema::PrimitiveType_DeDepthwiseConv2D] = PopulateDeconvDwParameter;
populate_parameter_funcs_[schema::PrimitiveType_DeConv2D] = PopulateDeconvParameter; populate_parameter_funcs_[schema::PrimitiveType_DeConv2D] = PopulateDeconvParameter;
populate_parameter_funcs_[schema::PrimitiveType_FusedBatchNorm] = PopulateFusedBatchNorm; populate_parameter_funcs_[schema::PrimitiveType_FusedBatchNorm] = PopulateFusedBatchNorm;
populate_parameter_funcs_[schema::PrimitiveType_BatchNorm] = PopulateBatchNorm;
populate_parameter_funcs_[schema::PrimitiveType_FullConnection] = PopulateFullconnectionParameter; populate_parameter_funcs_[schema::PrimitiveType_FullConnection] = PopulateFullconnectionParameter;
populate_parameter_funcs_[schema::PrimitiveType_Power] = PopulatePowerParameter; populate_parameter_funcs_[schema::PrimitiveType_Power] = PopulatePowerParameter;
populate_parameter_funcs_[schema::PrimitiveType_LocalResponseNormalization] = PopulateLocalResponseNormParameter; populate_parameter_funcs_[schema::PrimitiveType_LocalResponseNormalization] = PopulateLocalResponseNormParameter;
......
/**
* 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/batchnorm.h"
#include <cmath>
#include "schema/model_generated.h"
#include "src/kernel_registry.h"
#include "include/errorcode.h"
#include "src/runtime/runtime_api.h"
using mindspore::kernel::KERNEL_ARCH::kCPU;
using mindspore::lite::KernelRegistrar;
using mindspore::lite::RET_ERROR;
using mindspore::lite::RET_OK;
using mindspore::schema::PrimitiveType_BatchNorm;
namespace mindspore::kernel {
int BatchnormCPUKernel::Init() { return RET_OK; }
int BatchnormCPUKernel::ReSize() { return RET_OK; }
int BatchnormCPUKernel::DoExecute(int tid) {
int count = MSMIN(thread_unit_, units_ - tid * thread_unit_);
if (count <= 0) {
return RET_OK;
}
int offset = tid * thread_unit_ * channel_;
BatchNorm(in_addr_ + offset, mean_addr_, var_addr_, count, channel_, batchnorm_param_->epsilon_, out_addr_ + offset);
return RET_OK;
}
int BatchNormRun(int task_id, LiteParallelGroupEnv *penv, void *cdata) {
auto g_kernel = reinterpret_cast<BatchnormCPUKernel *>(cdata);
auto ret = g_kernel->DoExecute(task_id);
if (ret != RET_OK) {
MS_LOG(ERROR) << "BatchnormRun error task_id[" << task_id << "] error_code[" << ret << "]";
return ret;
}
return RET_OK;
}
int BatchnormCPUKernel::Run() {
in_addr_ = reinterpret_cast<float *>(inputs_.at(0)->Data());
mean_addr_ = reinterpret_cast<float *>(inputs_.at(1)->Data());
var_addr_ = reinterpret_cast<float *>(inputs_.at(2)->Data());
out_addr_ = reinterpret_cast<float *>(outputs_.at(0)->Data());
auto input_shapes = inputs_[0]->shape();
channel_ = input_shapes[3];
units_ = 1;
for (int i = 0; i < 3; i++) {
units_ *= input_shapes[i];
}
thread_count_ = MSMIN(thread_count_, units_);
thread_unit_ = UP_DIV(units_, thread_count_);
int ret = LiteBackendParallelLaunch(BatchNormRun, this, thread_count_);
if (ret != RET_OK) {
MS_LOG(ERROR) << "BatchnormRun error error_code[" << ret << "]";
return ret;
}
return RET_OK;
}
kernel::LiteKernel *CpuBatchnormKernelCreator(const std::vector<lite::tensor::Tensor *> &inputs,
const std::vector<lite::tensor::Tensor *> &outputs,
OpParameter *opParameter, const lite::Context *ctx,
const kernel::KernelKey &desc) {
MS_ASSERT(opParameter != nullptr);
MS_ASSERT(desc.type == schema::PrimitiveType_BatchNorm);
auto *kernel = new (std::nothrow) BatchnormCPUKernel(opParameter, inputs, outputs, ctx);
if (kernel == nullptr) {
MS_LOG(ERROR) << "new BatchNormCPUKernel fail!";
return nullptr;
}
auto ret = kernel->Init();
if (ret != RET_OK) {
MS_LOG(ERROR) << "Init kernel failed, name: " << opParameter->name_ << ", type: "
<< schema::EnumNamePrimitiveType(static_cast<schema::PrimitiveType>(opParameter->type_));
delete kernel;
return nullptr;
}
return kernel;
}
REG_KERNEL(kCPU, kNumberTypeFloat32, PrimitiveType_BatchNorm, CpuBatchnormKernelCreator)
} // namespace mindspore::kernel
/**
* 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_BATCHNORM_H_
#define MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_FP32_BATCHNORM_H_
#include <vector>
#include "src/lite_kernel.h"
#include "include/context.h"
#include "src/runtime/kernel/arm/opclib/fp32/batchnorm.h"
using mindspore::lite::Context;
namespace mindspore::kernel {
class BatchnormCPUKernel : public LiteKernel {
public:
BatchnormCPUKernel(OpParameter *parameter, const std::vector<lite::tensor::Tensor *> &inputs,
const std::vector<lite::tensor::Tensor *> &outputs, const Context *ctx)
: LiteKernel(parameter, inputs, outputs), ctx_(ctx), thread_count_(ctx->thread_num_) {
batchnorm_param_ = reinterpret_cast<BatchNormParameter *>(parameter);
}
~BatchnormCPUKernel() override { delete batchnorm_param_; }
int Init() override;
int ReSize() override;
int Run() override;
int DoExecute(int tid);
private:
int thread_count_;
int thread_unit_;
int units_;
int channel_;
float *in_addr_;
float *mean_addr_;
float *var_addr_;
float *out_addr_;
const Context *ctx_;
BatchNormParameter *batchnorm_param_;
};
} // namespace mindspore::kernel
#endif // MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_FP32_BATCHNORM_H_
/**
* 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/opclib/fp32/batchnorm.h"
void BatchNorm(const float *input_ptr, const float *mean_ptr, const float *variance_ptr, int units, int channel,
float epsilon, float *output_ptr) {
for (int u = 0; u < units; u++) {
for (int c = 0; c < channel; c++) {
auto variance_sqrt = sqrt(variance_ptr[c] + epsilon);
output_ptr[u * channel + c] = (input_ptr[u * channel + c] - mean_ptr[c]) / variance_sqrt;
}
}
}
/**
* 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_OPCLIB_FP32_BATCHNORM_H_
#define MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_OPCLIB_FP32_BATCHNORM_H_
#include "src/runtime/kernel/arm/opclib/op_base.h"
struct BatchNormParameter {
OpParameter op_parameter_;
float epsilon_;
};
void BatchNorm(const float *input_ptr, const float *mean_ptr, const float *variance_ptr, int count, int channel,
float epsilon, float *output_ptr);
#endif // MINDSPORE_LITE_SRC_RUNTIME_KERNEL_ARM_OPCLIB_FUSED_BATCHNORM_H_
...@@ -96,8 +96,8 @@ MetaGraphTptr BuildCaffeGraph(schema::PrimitiveType conv_type) { ...@@ -96,8 +96,8 @@ MetaGraphTptr BuildCaffeGraph(schema::PrimitiveType conv_type) {
bn_node->inputIndex = {2, 3, 4}; bn_node->inputIndex = {2, 3, 4};
bn_node->outputIndex = {5}; bn_node->outputIndex = {5};
bn_node->primitive = std::make_unique<schema::PrimitiveT>(); bn_node->primitive = std::make_unique<schema::PrimitiveT>();
bn_node->primitive->value.type = schema::PrimitiveType_CaffeBatchNorm; bn_node->primitive->value.type = schema::PrimitiveType_BatchNorm;
auto prim2 = new schema::CaffeBatchNormT; auto prim2 = new schema::BatchNormT;
bn_node->primitive->value.value = prim2; bn_node->primitive->value.value = prim2;
bn_node->name = "bn"; bn_node->name = "bn";
meta_graph->nodes.emplace_back(std::move(bn_node)); meta_graph->nodes.emplace_back(std::move(bn_node));
......
/**
* 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 <iostream>
#include "mindspore/core/utils/log_adapter.h"
#include "common/common_test.h"
#include "mindspore/lite/src/runtime/kernel/arm/opclib/fp32/batchnorm.h"
#include "mindspore/lite/src/runtime/kernel/arm/opclib/fused_batchnorm.h"
#include "mindspore/lite/src/kernel_registry.h"
#include "mindspore/lite/src/lite_kernel.h"
#include "mindspore/lite/src/common/file_utils.h"
namespace mindspore {
class TestBatchnormFp32 : public mindspore::Common {
public:
TestBatchnormFp32() {}
};
TEST_F(TestBatchnormFp32, BNTest) {
std::vector<float> in_data = {0.0669681, 0.959215, 0.252686, 0.613594, 0.811776, 0.139469, 0.322848, 0.118354,
0.082978, 0.399467, 0.961267, 0.0247456, 0.0714259, 0.0791484, 0.0648625, 0.561612,
0.412069, 0.311492, 0.46109, 0.377125, 0.369283, 0.0332446, 0.696142, 0.715973,
0.525524, 0.477265, 0.0336351, 0.751577, 0.377548, 0.964603, 0.0196834, 0.174865};
std::vector<float> in_data1 = {0.855446, 0.821765, 0.281008, 0.0798653, 0.22294, 0.793782, 0.963222, 0.17851,
0.667549, 0.274381, 0.592842, 0.216552, 0.190274, 0.237873, 0.610063, 0.307559,
0.830007, 0.760957, 0.583265, 0.763793, 0.456372, 0.391378, 0.547915, 0.862198,
0.510794, 0.826776, 0.515894, 0.30071, 0.404987, 0.184773};
std::vector<float> in_data2 = {0.712438, 0.4927, 0.078419, 0.310429, 0.546871, 0.0667141, 0.874321, 0.0265647,
0.685165, 0.732586, 0.952889, 0.506402, 0.540784, 0.131119, 0.357713, 0.678992,
0.960839, 0.340706, 0.697678, 0.398146, 0.313321, 0.6485, 0.739153, 0.00190134,
0.536842, 0.996873, 0.445276, 0.371212, 0.420397, 0.0930115};
std::vector<float> in_data3(32, 1);
std::vector<float> in_data4(32, 0);
std::vector<lite::tensor::Tensor *> inputs_tensor;
std::vector<lite::tensor::Tensor *> outputs_tensor;
BatchNormParameter op_param;
op_param.op_parameter_.type_ = schema::PrimitiveType_BatchNorm;
op_param.epsilon_ = 0.001f;
std::vector<int> in_shape = {1, 2, 4, 4};
lite::tensor::Tensor input0_tensor;
lite::tensor::Tensor input1_tensor;
lite::tensor::Tensor input2_tensor;
inputs_tensor.push_back(&input0_tensor);
inputs_tensor.push_back(&input1_tensor);
inputs_tensor.push_back(&input2_tensor);
input0_tensor.SetData(in_data.data());
input1_tensor.SetData(in_data1.data());
input2_tensor.SetData(in_data2.data());
input0_tensor.set_shape(in_shape);
std::vector<float> output(32);
std::vector<float> corr_out(32);
std::vector<int> output_shape = {1, 2, 4, 4};
lite::tensor::Tensor output0_tensor;
outputs_tensor.push_back(&output0_tensor);
output0_tensor.SetData(output.data());
kernel::KernelKey desc = {kernel::KERNEL_ARCH::kCPU, kNumberTypeFloat32, schema::PrimitiveType_BatchNorm};
auto creator = lite::KernelRegistry::GetInstance()->GetCreator(desc);
ASSERT_NE(creator, nullptr);
lite::Context ctx;
ctx.thread_num_ = 7;
kernel::LiteKernel *kernel =
creator(inputs_tensor, outputs_tensor, reinterpret_cast<OpParameter *>(&op_param), &ctx, desc);
ASSERT_NE(kernel, nullptr);
auto output_tensor_shape = output0_tensor.shape();
kernel->Run();
FusedBatchNorm(in_data.data(), in_data3.data(), in_data4.data(), in_data1.data(), in_data2.data(), in_shape.data(),
0.001f, corr_out.data());
printf("==================output data=================\n");
for (int i = 0; i < 1 * 28; i++) {
std::cout << output[i] << " ,";
}
std::cout << std::endl;
CompareOutputData(output.data(), corr_out.data(), 32, 0.00001);
input0_tensor.SetData(nullptr);
input1_tensor.SetData(nullptr);
input2_tensor.SetData(nullptr);
output0_tensor.SetData(nullptr);
}
} // namespace mindspore
...@@ -50,7 +50,7 @@ STATUS ConvBNFusionPass::DefinePattern() { ...@@ -50,7 +50,7 @@ STATUS ConvBNFusionPass::DefinePattern() {
convOp->types = {schema::PrimitiveType_Conv2D, schema::PrimitiveType_DepthwiseConv2D}; convOp->types = {schema::PrimitiveType_Conv2D, schema::PrimitiveType_DepthwiseConv2D};
auto bnOp = std::make_shared<PatternOp>(); auto bnOp = std::make_shared<PatternOp>();
bnOp->id = DST_NAME; bnOp->id = DST_NAME;
bnOp->types = {schema::PrimitiveType_FusedBatchNorm, schema::PrimitiveType_CaffeBatchNorm}; bnOp->types = {schema::PrimitiveType_FusedBatchNorm, schema::PrimitiveType_BatchNorm};
bnOp->left = convOp; bnOp->left = convOp;
std::unique_ptr<FusionPattern> fusionPattern(new (std::nothrow) FusionPattern("ConvBatchNormFusion")); std::unique_ptr<FusionPattern> fusionPattern(new (std::nothrow) FusionPattern("ConvBatchNormFusion"));
...@@ -208,8 +208,8 @@ STATUS ConvBNFusionPass::GetBnEpsilon(schema::MetaGraphT *graph, std::shared_ptr ...@@ -208,8 +208,8 @@ STATUS ConvBNFusionPass::GetBnEpsilon(schema::MetaGraphT *graph, std::shared_ptr
MS_ASSERT(bnNode != nullptr); MS_ASSERT(bnNode != nullptr);
if (bnNode->primitive->value.type == schema::PrimitiveType_FusedBatchNorm) { if (bnNode->primitive->value.type == schema::PrimitiveType_FusedBatchNorm) {
eps = bnNode->primitive->value.AsFusedBatchNorm()->epsilon; eps = bnNode->primitive->value.AsFusedBatchNorm()->epsilon;
} else if (bnNode->primitive->value.type == schema::PrimitiveType_CaffeBatchNorm) { } else if (bnNode->primitive->value.type == schema::PrimitiveType_BatchNorm) {
eps = bnNode->primitive->value.AsCaffeBatchNorm()->epsilon; eps = bnNode->primitive->value.AsBatchNorm()->epsilon;
} else { } else {
MS_LOG(ERROR) << "match pattern has error, " << bnNode->name.c_str() << " not BatchNorm node"; MS_LOG(ERROR) << "match pattern has error, " << bnNode->name.c_str() << " not BatchNorm node";
return RET_ERROR; return RET_ERROR;
......
...@@ -28,13 +28,11 @@ static const int CAFFE_BATCHNORMAL_TOP_SIZE = 1; ...@@ -28,13 +28,11 @@ static const int CAFFE_BATCHNORMAL_TOP_SIZE = 1;
namespace mindspore { namespace mindspore {
namespace lite { namespace lite {
using STATUS = int; using STATUS = int;
STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto, STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto, const caffe::LayerParameter &weight,
const caffe::LayerParameter &weight, schema::CNodeT *op, std::vector<schema::TensorT *> *weightVec) {
schema::CNodeT *op,
std::vector<schema::TensorT *> *weightVec) {
op->name = proto.name(); op->name = proto.name();
// caffe batch norm attr // caffe batch norm attr
std::unique_ptr<FusedBatchNormT> attr(new FusedBatchNormT()); std::unique_ptr<schema::BatchNormT> attr(new schema::BatchNormT());
const caffe::BatchNormParameter batchNormParam = proto.batch_norm_param(); const caffe::BatchNormParameter batchNormParam = proto.batch_norm_param();
// check bottom size // check bottom size
...@@ -98,7 +96,7 @@ STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto, ...@@ -98,7 +96,7 @@ STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto,
weightVec->push_back(beta); weightVec->push_back(beta);
op->primitive = std::make_unique<schema::PrimitiveT>(); op->primitive = std::make_unique<schema::PrimitiveT>();
op->primitive->value.type = schema::PrimitiveType_FusedBatchNorm; op->primitive->value.type = schema::PrimitiveType_BatchNorm;
op->primitive->value.value = attr.release(); op->primitive->value.value = attr.release();
return RET_OK; return RET_OK;
...@@ -107,5 +105,3 @@ STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto, ...@@ -107,5 +105,3 @@ STATUS CaffeBatchNormParser::Parse(const caffe::LayerParameter &proto,
CaffeNodeRegistrar g_caffeBatchNormParser("BatchNorm", new CaffeBatchNormParser()); CaffeNodeRegistrar g_caffeBatchNormParser("BatchNorm", new CaffeBatchNormParser());
} // namespace lite } // namespace lite
} // namespace mindspore } // namespace mindspore
...@@ -88,14 +88,13 @@ schema::MetaGraphT *CaffeModelParser::Parse(const std::string &modelFile, const ...@@ -88,14 +88,13 @@ schema::MetaGraphT *CaffeModelParser::Parse(const std::string &modelFile, const
SetAllTensors(tensorCache, subGraphDef.get()); SetAllTensors(tensorCache, subGraphDef.get());
graph = move(subGraphDef); graph = move(subGraphDef);
ConvertCaffeBatchNorm(graph.get()); // ConvertCaffeBatchNorm(graph.get());
return graph.release(); return graph.release();
// return Fb2Anf(graph.release()); // return Fb2Anf(graph.release());
} }
STATUS CaffeModelParser::SetOpInputIdx(const caffe::LayerParameter &layer, STATUS CaffeModelParser::SetOpInputIdx(const caffe::LayerParameter &layer, schema::CNodeT *op,
schema::CNodeT *op,
TensorCache *tensorCache) { TensorCache *tensorCache) {
for (int i = 0; i < layer.bottom_size(); i++) { for (int i = 0; i < layer.bottom_size(); i++) {
int index = tensorCache->FindTensor(layer.bottom(i)); int index = tensorCache->FindTensor(layer.bottom(i));
...@@ -109,8 +108,7 @@ STATUS CaffeModelParser::SetOpInputIdx(const caffe::LayerParameter &layer, ...@@ -109,8 +108,7 @@ STATUS CaffeModelParser::SetOpInputIdx(const caffe::LayerParameter &layer,
return RET_OK; return RET_OK;
} }
STATUS CaffeModelParser::SetOpOutputIdx(const caffe::LayerParameter &layer, STATUS CaffeModelParser::SetOpOutputIdx(const caffe::LayerParameter &layer, schema::CNodeT *op,
schema::CNodeT *op,
TensorCache *tensorCache) { TensorCache *tensorCache) {
for (int i = 0; i < layer.top_size(); i++) { for (int i = 0; i < layer.top_size(); i++) {
std::unique_ptr<schema::TensorT> msTensor(new schema::TensorT()); std::unique_ptr<schema::TensorT> msTensor(new schema::TensorT());
...@@ -279,7 +277,7 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) { ...@@ -279,7 +277,7 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) {
scaleTensor->dataType = TypeId::kNumberTypeFloat32; scaleTensor->dataType = TypeId::kNumberTypeFloat32;
scaleTensor->data.resize(shapeSize * sizeof(float)); scaleTensor->data.resize(shapeSize * sizeof(float));
auto scaleData = reinterpret_cast<float *>(scaleTensor->data.data()); auto scaleData = reinterpret_cast<float *>(scaleTensor->data.data());
for (size_t i = 0 ; i < shapeSize; i++) { for (size_t i = 0; i < shapeSize; i++) {
scaleData[i] = 1; scaleData[i] = 1;
} }
...@@ -291,7 +289,7 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) { ...@@ -291,7 +289,7 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) {
biasTensor->dataType = TypeId::kNumberTypeInt32; biasTensor->dataType = TypeId::kNumberTypeInt32;
biasTensor->data.resize(shapeSize * sizeof(int32_t)); biasTensor->data.resize(shapeSize * sizeof(int32_t));
auto biasData = reinterpret_cast<int32_t *>(biasTensor->data.data()); auto biasData = reinterpret_cast<int32_t *>(biasTensor->data.data());
for (size_t i = 0 ; i < shapeSize; i++) { for (size_t i = 0; i < shapeSize; i++) {
biasData[i] = 0; biasData[i] = 0;
} }
...@@ -304,4 +302,3 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) { ...@@ -304,4 +302,3 @@ void CaffeModelParser::ConvertCaffeBatchNorm(schema::MetaGraphT *meta_graph) {
} }
} // namespace lite } // namespace lite
} // namespace mindspore } // namespace mindspore
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册