【Caffe代码解析】Layer网络层

Layer 功能：

是所有的网络层的基类，其中，定义了一些通用的接口，比如前馈，反馈，reshape，setup等。

#ifndef CAFFE_LAYER_H_
#define CAFFE_LAYER_H_#include <algorithm>
#include <string>
#include <vector>#include "caffe/blob.hpp"
#include "caffe/common.hpp"
#include "caffe/layer_factory.hpp"
#include "caffe/proto/caffe.pb.h"
#include "caffe/util/device_alternate.hpp"namespace caffe {// 功能：所有的网络层的基类，定义的所有的网络层的通用接口。// 前馈接口，必须实现// 反馈接口，需要的时候实现，计算梯度。
template <typename Dtype>
class Layer {public:/*** 每个网络层需要自己定义它的setup而不需要构造函数*/explicit Layer(const LayerParameter& param): layer_param_(param) {//通过网络层参数来构造网络层phase_ = param.phase();if (layer_param_.blobs_size() > 0) {blobs_.resize(layer_param_.blobs_size());for (int i = 0; i < layer_param_.blobs_size(); ++i) {blobs_[i].reset(new Blob<Dtype>());blobs_[i]->FromProto(layer_param_.blobs(i));}}}// 析构函数virtual ~Layer() {}/*** 实现一些通用的设置功能** @param bottom 网络层的输入的shape* @param top 网络层的输出，需要被reshape* 调用 LayerSetUp 来对每一个网络层进行特殊化的处理,* 调用reshape top* 设置 数值权重* 这个方法可以不被重载。*/void SetUp(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) {CheckBlobCounts(bottom, top);LayerSetUp(bottom, top);Reshape(bottom, top);SetLossWeights(top);}/*** @brief 设置一些层相关的设置，定义的层需要实现这个方法以及Reshape方法*///设置网络层virtual void LayerSetUp(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) {}/*** @brief 调整top blob以适应bottom blob。*/virtual void Reshape(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) = 0;/*** @brief 给定 bottom blobs, 计算 top blobs 以及 loss.* 每一个网络层都需要定义cpu版本的前馈，可选gpu版本的前馈*///前馈inline Dtype Forward(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top);/*** @brief 给定 top blob 的梯度, 计算 bottom blob 梯度.* @param propagate_down 向量，长度为ibottom   的个数，每个索引值表示是是否将损失梯度值反馈到该bottom中*///反馈inline void Backward(const vector<Blob<Dtype>*>& top,const vector<bool>& propagate_down,const vector<Blob<Dtype>*>& bottom);/*** @brief 返回可学习的参数 blobs.*/vector<shared_ptr<Blob<Dtype> > >& blobs() {return blobs_;}/*** @brief 返回网络层参数*/const LayerParameter& layer_param() const { return layer_param_; }//序列化virtual void ToProto(LayerParameter* param, bool write_diff = false);/*** @brief 返回指定索引的标量损失值。*/inline Dtype loss(const int top_index) const {return (loss_.size() > top_index) ? loss_[top_index] : Dtype(0);}/*** @brief 设置网络层制定索引位置的loss*/inline void set_loss(const int top_index, const Dtype value) {if (loss_.size() <= top_index) {loss_.resize(top_index + 1, Dtype(0));}loss_[top_index] = value;}/*** @brief 返回网络层名字，字符串描述u*/virtual inline const char* type() const { return ""; }//Bottom的blob的确切数目virtual inline int ExactNumBottomBlobs() const { return -1; }//Bottom blob的最小数目virtual inline int MinBottomBlobs() const { return -1; }//Botttom的确切数目virtual inline int MaxBottomBlobs() const { return -1; }//Top Blob的确切数目virtual inline int ExactNumTopBlobs() const { return -1; }//最小的blob的数目virtual inline int MinTopBlobs() const { return -1; }// 最大的blob的数目virtual inline int MaxTopBlobs() const { return -1; }// 是否bottom 和top的数目相同virtual inline bool EqualNumBottomTopBlobs() const { return false; }// 是否自动Top blobvirtual inline bool AutoTopBlobs() const { return false; }//查询某一个bottom是否强制bpvirtual inline bool AllowForceBackward(const int bottom_index) const {return true;}//查询某一个blob是否bpinline bool param_propagate_down(const int param_id) {return (param_propagate_down_.size() > param_id) ?param_propagate_down_[param_id] : false;}//设置某一个blob是否bp。inline void set_param_propagate_down(const int param_id, const bool value) {if (param_propagate_down_.size() <= param_id) {param_propagate_down_.resize(param_id + 1, true);}param_propagate_down_[param_id] = value;}protected:// 网络层参数LayerParameter layer_param_;// 模式Phase phase_;//用blob来存储一系列向量vector<shared_ptr<Blob<Dtype> > > blobs_;//是否bp的向量vector<bool> param_propagate_down_;//存储top的lossvector<Dtype> loss_;//cpu版本的前馈实现virtual void Forward_cpu(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) = 0;//gpu版本的前馈实现virtual void Forward_gpu(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) {// LOG(WARNING) << "Using CPU code as backup.";return Forward_cpu(bottom, top);}//cpu版本的前馈实现virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,const vector<bool>& propagate_down,const vector<Blob<Dtype>*>& bottom) = 0;//gpu版本的反馈实现virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,const vector<bool>& propagate_down,const vector<Blob<Dtype>*>& bottom) {// LOG(WARNING) << "Using CPU code as backup.";Backward_cpu(top, propagate_down, bottom);}// 核查bootom和top的大小是否与该layer层指定的一致。virtual void CheckBlobCounts(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) {if (ExactNumBottomBlobs() >= 0) {CHECK_EQ(ExactNumBottomBlobs(), bottom.size())<< type() << " Layer takes " << ExactNumBottomBlobs()<< " bottom blob(s) as input.";}if (MinBottomBlobs() >= 0) {CHECK_LE(MinBottomBlobs(), bottom.size())<< type() << " Layer takes at least " << MinBottomBlobs()<< " bottom blob(s) as input.";}if (MaxBottomBlobs() >= 0) {CHECK_GE(MaxBottomBlobs(), bottom.size())<< type() << " Layer takes at most " << MaxBottomBlobs()<< " bottom blob(s) as input.";}if (ExactNumTopBlobs() >= 0) {CHECK_EQ(ExactNumTopBlobs(), top.size())<< type() << " Layer produces " << ExactNumTopBlobs()<< " top blob(s) as output.";}if (MinTopBlobs() >= 0) {CHECK_LE(MinTopBlobs(), top.size())<< type() << " Layer produces at least " << MinTopBlobs()<< " top blob(s) as output.";}if (MaxTopBlobs() >= 0) {CHECK_GE(MaxTopBlobs(), top.size())<< type() << " Layer produces at most " << MaxTopBlobs()<< " top blob(s) as output.";}if (EqualNumBottomTopBlobs()) {CHECK_EQ(bottom.size(), top.size())<< type() << " Layer produces one top blob as output for each "<< "bottom blob input.";}}// 用blob初始化损失权重。inline void SetLossWeights(const vector<Blob<Dtype>*>& top) {const int num_loss_weights = layer_param_.loss_weight_size();if (num_loss_weights) {CHECK_EQ(top.size(), num_loss_weights) << "loss_weight must be ""unspecified or specified once per top blob.";for (int top_id = 0; top_id < top.size(); ++top_id) {const Dtype loss_weight = layer_param_.loss_weight(top_id);if (loss_weight == Dtype(0)) { continue; }this->set_loss(top_id, loss_weight);const int count = top[top_id]->count();Dtype* loss_multiplier = top[top_id]->mutable_cpu_diff();caffe_set(count, loss_weight, loss_multiplier);}}}DISABLE_COPY_AND_ASSIGN(Layer);
};  // class Layer// 前馈，根据caffe的mode 调用相对应的cpu实现或者是gpu实现，并且计算损失函数值。
template <typename Dtype>
inline Dtype Layer<Dtype>::Forward(const vector<Blob<Dtype>*>& bottom,const vector<Blob<Dtype>*>& top) {Dtype loss = 0;Reshape(bottom, top);switch (Caffe::mode()) {case Caffe::CPU:Forward_cpu(bottom, top);for (int top_id = 0; top_id < top.size(); ++top_id) {if (!this->loss(top_id)) { continue; }const int count = top[top_id]->count();const Dtype* data = top[top_id]->cpu_data();const Dtype* loss_weights = top[top_id]->cpu_diff();loss += caffe_cpu_dot(count, data, loss_weights);}break;case Caffe::GPU:Forward_gpu(bottom, top);
#ifndef CPU_ONLYfor (int top_id = 0; top_id < top.size(); ++top_id) {if (!this->loss(top_id)) { continue; }const int count = top[top_id]->count();const Dtype* data = top[top_id]->gpu_data();const Dtype* loss_weights = top[top_id]->gpu_diff();Dtype blob_loss = 0;caffe_gpu_dot(count, data, loss_weights, &blob_loss);loss += blob_loss;}
#endifbreak;default:LOG(FATAL) << "Unknown caffe mode.";}return loss;
}//反向传播梯度，根据Caffe的mode是在GPU还是CPU，调用相对应版本的函数
//propagate_down 用于控制对应的层是否bp
template <typename Dtype>
inline void Layer<Dtype>::Backward(const vector<Blob<Dtype>*>& top,const vector<bool>& propagate_down,const vector<Blob<Dtype>*>& bottom) {switch (Caffe::mode()) {case Caffe::CPU:Backward_cpu(top, propagate_down, bottom);break;case Caffe::GPU:Backward_gpu(top, propagate_down, bottom);break;default:LOG(FATAL) << "Unknown caffe mode.";}
}// 序列化网络层参数到协议缓存，最终是调用blob写入协议缓存。
template <typename Dtype>
void Layer<Dtype>::ToProto(LayerParameter* param, bool write_diff) {param->Clear();param->CopyFrom(layer_param_);param->clear_blobs();for (int i = 0; i < blobs_.size(); ++i) {blobs_[i]->ToProto(param->add_blobs(), write_diff);}
}}  // namespace caffe#endif  // CAFFE_LAYER_H_

【Caffe代码解析】Layer网络层相关推荐

梳理caffe代码layer(五)
Layer(层)是Caffe中最庞大最繁杂的模块.由于Caffe强调模块化设计,因此只允许每个layer完成一类特定的计算,例如convolution操作.pooling.非线性变换.内积运算,以及数 ...
EfficientNet（ICML 2019）原理与代码解析
paper:EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks code:mmclassification ...
DeepLearning | 图注意力网络Graph Attention Network（GAT）论文、模型、代码解析
本篇博客是对论文 Velikovi, Petar, Cucurull, Guillem, Casanova, Arantxa,et al. Graph Attention Networks, 2018 ...
Caffe代码导读（5）：对数据集进行Testing
转载自: Caffe代码导读(5):对数据集进行Testing - 卜居 - 博客频道 - CSDN.NET http://blog.csdn.net/kkk584520/article/detail ...
Caffe代码导读（0）：路线图
转载自: Caffe代码导读(0):路线图 - 卜居 - 博客频道 - CSDN.NET http://blog.csdn.net/kkk584520/article/details/41681085 ...
Python编写caffe代码
有时候,我们需要将网络使用caffe代码实现,人工手写容易出问题.可以使用Python完成网络编写. 卷积层: def generate_conv_layer_no_bias(name, bottom ...
[GCN] 代码解析 of GitHub：Semi-supervised classification with graph convolutional networks
本文解析的代码是论文Semi-Supervised Classification with Graph Convolutional Networks作者提供的实现代码. 原GitHub:Graph C ...
Caffe编写Python layer
Caffe编写Python layer 在使用caffe做训练的时候,通常的做法是把数据转为lmdb格式,然后在train.prototxt中指定,最后在开始训练,转为lmdb格式的优点是读取数据更高 ...
Rescue-Prime hash STARK 代码解析
1. 引言前序博客有: STARK入门知识 STARKs and STARK VM: Proofs of Computational Integrity STARK中的FRI代码解析 Rescue- ...
Polygon zkEVM的pil-stark Fibonacci状态机代码解析
1. 引言前序博客有: Polygon zkEVM的pil-stark Fibonacci状态机初体验 STARKs and STARK VM: Proofs of Computational In ...

【Caffe代码解析】Layer网络层

【Caffe代码解析】Layer网络层相关推荐

最新文章

热门文章