Caffe源码-Layer类

Posted 2020-11-24 relu110

Layer类简介

Layer是caffe中搭建网络的基本单元，caffe代码中包含大量Layer基类派生出来的各种各样的层，各自通过虚函数 Forward() 和 Backward() 实现自己的功能。 Forward() 函数用于前向计算过程，由 bottom blob 计算 top blob 和 loss ，实现数据由浅至深的传递。而 Backward() 函数用于反向传播过程，由 top blob 的计算 bottom blob 的梯度，将网络的预测误差向浅层网络传递，以便更新网络的参数。

layer.hpp源码

/**
 * @brief An interface for the units of computation which can be composed into a
 *        Net.
 *
 * Layer%s must implement a Forward function, in which they take their input
 * (bottom) Blob%s (if any) and compute their output Blob%s (if any).
 * They may also implement a Backward function, in which they compute the error
 * gradients with respect to their input Blob%s, given the error gradients with
 * their output Blob%s.
 */
template <typename Dtype>
class Layer {
 public:
  /**
   * You should not implement your own constructor. Any set up code should go
   * to SetUp(), where the dimensions of the bottom blobs are provided to the
   * layer.
   */
  explicit Layer(const LayerParameter& param)       //构造函数,根据LayerParameter初始化当前的layer
    : layer_param_(param) {                         //先保存param
      // Set phase and copy blobs (if there are any).
      phase_ = param.phase();                       //设置layer的状态(TRAIN or TEST)
      if (layer_param_.blobs_size() > 0) {          //如果layer中存在blob参数(可学习参数)
        blobs_.resize(layer_param_.blobs_size());   //调整blobs_的大小
        for (int i = 0; i < layer_param_.blobs_size(); ++i) {
          blobs_[i].reset(new Blob<Dtype>());       //指针指向新创建的Blob数据
          blobs_[i]->FromProto(layer_param_.blobs(i));    //读取layer_param_中的BlobProto类型的数据,存入blob中
        }
      }
    }
  virtual ~Layer() {}

  /**
   * @brief Implements common layer setup functionality.
   *
   * @param bottom the preshaped input blobs
   * @param top
   *     the allocated but unshaped output blobs, to be shaped by Reshape
   *
   * Checks that the number of bottom and top blobs is correct.
   * Calls LayerSetUp to do special layer setup for individual layer types,
   * followed by Reshape to set up sizes of top blobs and internal buffers.
   * Sets up the loss weight multiplier blobs for any non-zero loss weights.
   * This method may not be overridden.
   */
  void SetUp(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) {
    CheckBlobCounts(bottom, top);   //检查输入输出的blob的个数是否正确
    LayerSetUp(bottom, top);        //虚函数,layer创建时做一些初始化工作,每个子类根据需要自己实现
    Reshape(bottom, top);           //虚函数,调整输出blob的形状和layer中内部缓冲区的大小
    SetLossWeights(top);            //设置输出blob的权重
  }

  /**
   * @brief Does layer-specific setup: your layer should implement this function
   *        as well as Reshape.
   *
   * @param bottom
   *     the preshaped input blobs, whose data fields store the input data for
   *     this layer
   * @param top
   *     the allocated but unshaped output blobs
   *
   * This method should do one-time layer specific setup. This includes reading
   * and processing relevent parameters from the <code>layer_param_</code>.
   * Setting up the shapes of top blobs and internal buffers should be done in
   * <code>Reshape</code>, which will be called before the forward pass to
   * adjust the top blob sizes.
   */
  //特定层有特定的实现方式,一般用于在layer创建时做一些一次性的操作,如从layer_param_中读取点参数之类的
  virtual void LayerSetUp(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) {}

  /**
   * @brief Adjust the shapes of top blobs and internal buffers to accommodate
   *        the shapes of the bottom blobs.
   *
   * @param bottom the input blobs, with the requested input shapes
   * @param top the top blobs, which should be reshaped as needed
   *
   * This method should reshape top blobs as needed according to the shapes
   * of the bottom (input) blobs, as well as reshaping any internal buffers
   * and making any other necessary adjustments so that the layer can
   * accommodate the bottom blobs.
   */
  //根据输入blob的形状,调整输出blob的形状和调整内部的缓冲区.该函数会在每次前向计算之前被调用
  virtual void Reshape(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) = 0;

  /**
   * @brief Given the bottom blobs, compute the top blobs and the loss.
   *
   * @param bottom
   *     the input blobs, whose data fields store the input data for this layer
   * @param top
   *     the preshaped output blobs, whose data fields will store this layers'
   *     outputs
   * 
eturn The total loss from the layer.
   *
   * The Forward wrapper calls the relevant device wrapper function
   * (Forward_cpu or Forward_gpu) to compute the top blob values given the
   * bottom blobs.  If the layer has any non-zero loss_weights, the wrapper
   * then computes and returns the loss.
   *
   * Your layer should implement Forward_cpu and (optionally) Forward_gpu.
   */
  //前向计算函数,传入输入blob数据,计算输出blob,和loss值(权重参数非0时).
  inline Dtype Forward(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top);

  /**
   * @brief Given the top blob error gradients, compute the bottom blob error
   *        gradients.
   *
   * @param top
   *     the output blobs, whose diff fields store the gradient of the error
   *     with respect to themselves
   * @param propagate_down
   *     a vector with equal length to bottom, with each index indicating
   *     whether to propagate the error gradients down to the bottom blob at
   *     the corresponding index
   * @param bottom
   *     the input blobs, whose diff fields will store the gradient of the error
   *     with respect to themselves after Backward is run
   *
   * The Backward wrapper calls the relevant device wrapper function
   * (Backward_cpu or Backward_gpu) to compute the bottom blob diffs given the
   * top blob diffs.
   *
   * Your layer should implement Backward_cpu and (optionally) Backward_gpu.
   */
  //给出输出blob数据的误差梯度,计算输入blob数据的误差梯度
  inline void Backward(const vector<Blob<Dtype>*>& top,
      const vector<bool>& propagate_down,   //个数与bottom.size()相等,对应每个输入blob是否需要反向传播
      const vector<Blob<Dtype>*>& bottom);

  /**
   * @brief Returns the vector of learnable parameter blobs.
   */
  vector<shared_ptr<Blob<Dtype> > >& blobs() {    //返回layer中内部可学习的参数的个数
    return blobs_;
  }

  /**
   * @brief Returns the layer parameter.
   */
  const LayerParameter& layer_param() const { return layer_param_; }      //返回layer的配置参数

  /**
   * @brief Writes the layer parameter to a protocol buffer
   */
  virtual void ToProto(LayerParameter* param, bool write_diff = false);   //将layer的blob数据写入param中

  /**
   * @brief Returns the scalar loss associated with a top blob at a given index.
   */
  inline Dtype loss(const int top_index) const {
    return (loss_.size() > top_index) ? loss_[top_index] : Dtype(0);    //返回第top_index个输出blob对应的loss权重
  }

  /**
   * @brief Sets the loss associated with a top blob at a given index.
   */
  inline void set_loss(const int top_index, const Dtype value) {    //设置第top_index个输出blob对应的loss的权重为value
    if (loss_.size() <= top_index) {
      loss_.resize(top_index + 1, Dtype(0));    //个数小于top_index,则调整大小,并用Dtype(0)初始化一下
    }
    loss_[top_index] = value;     //对应的loss权重设置为value
  }

  /**
   * @brief Returns the layer type.
   */
  virtual inline const char* type() const { return ""; }    //返回layer的类型

  /**
   * @brief Returns the exact number of bottom blobs required by the layer,
   *        or -1 if no exact number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some exact number of bottom blobs.
   */
  virtual inline int ExactNumBottomBlobs() const { return -1; }   //layer中要求输入blob的准确个数,-1表示无要求
  /**
   * @brief Returns the minimum number of bottom blobs required by the layer,
   *        or -1 if no minimum number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some minimum number of bottom blobs.
   */
  virtual inline int MinBottomBlobs() const { return -1; }        //layer中要求输入blob的最小个数,-1表示无要求
  /**
   * @brief Returns the maximum number of bottom blobs required by the layer,
   *        or -1 if no maximum number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some maximum number of bottom blobs.
   */
  virtual inline int MaxBottomBlobs() const { return -1; }        //layer中要求输入blob的最大个数,-1表示无要求
  /**
   * @brief Returns the exact number of top blobs required by the layer,
   *        or -1 if no exact number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some exact number of top blobs.
   */
  virtual inline int ExactNumTopBlobs() const { return -1; }      //layer中要求输出blob的准确个数,-1表示无要求
  /**
   * @brief Returns the minimum number of top blobs required by the layer,
   *        or -1 if no minimum number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some minimum number of top blobs.
   */
  virtual inline int MinTopBlobs() const { return -1; }           //layer中要求输出blob的最小个数,-1表示无要求
  /**
   * @brief Returns the maximum number of top blobs required by the layer,
   *        or -1 if no maximum number is required.
   *
   * This method should be overridden to return a non-negative value if your
   * layer expects some maximum number of top blobs.
   */
  virtual inline int MaxTopBlobs() const { return -1; }           //layer中要求输出blob的最大个数,-1表示无要求
  /**
   * @brief Returns true if the layer requires an equal number of bottom and
   *        top blobs.
   *
   * This method should be overridden to return true if your layer expects an
   * equal number of bottom and top blobs.
   */
  virtual inline bool EqualNumBottomTopBlobs() const { return false; }  //layer中要求输入blob的个数与输出blob的个数是否相等

  /**
   * @brief Return whether "anonymous" top blobs are created automatically
   *        by the layer.
   *
   * If this method returns true, Net::Init will create enough "anonymous" top
   * blobs to fulfill the requirement specified by ExactNumTopBlobs() or
   * MinTopBlobs().
   */
  //是否允许自动创建匿名blob.是则在Net::Init()中会自动创建匿名输出blob,直至其个数达到ExactNumTopBlobs()和MinTopBlobs()的最大值
  virtual inline bool AutoTopBlobs() const { return false; }    //TODO 匿名blob的用途暂时还不了解

  /**
   * @brief Return whether to allow force_backward for a given bottom blob
   *        index.
   *
   * If AllowForceBackward(i) == false, we will ignore the force_backward
   * setting and backpropagate to blob i only if it needs gradient information
   * (as is done when force_backward == false).
   */
  //允许的话则当net设置了强制反传时,layer的每个输出blob都必须遵从net的设置.如果不允许的话,则layer的输出blob是否需要反向传播
  //仍是根据自身的设置来决定,不必考虑net的设置.具体可见 net.cpp -> Init() -> if(param.force_backward())...
  virtual inline bool AllowForceBackward(const int bottom_index) const {    //layer是否允许强制反向传播
    return true;
  }

  /**
   * @brief Specifies whether the layer should compute gradients w.r.t. a
   *        parameter at a particular index given by param_id.
   *
   * You can safely ignore false values and always compute gradients
   * for all parameters, but possibly with wasteful computation.
   */
  inline bool param_propagate_down(const int param_id) {  //返回layer中第param_id个参数blob是否需要计算梯度
    return (param_propagate_down_.size() > param_id) ?
        param_propagate_down_[param_id] : false;          //超出param_id返回为false
  }
  /**
   * @brief Sets whether the layer should compute gradients w.r.t. a
   *        parameter at a particular index given by param_id.
   */
  inline void set_param_propagate_down(const int param_id, const bool value) {  //设置layer中第param_id个参数blob是否需要计算梯度
    if (param_propagate_down_.size() <= param_id) {
      param_propagate_down_.resize(param_id + 1, true);   //调整大小
    }
    param_propagate_down_[param_id] = value;    //设置值为value
  }

 protected:
  /** The protobuf that stores the layer parameters */
  //layer_param_中存放着layer中的各种参数,不过使用了protobuf库,例如layer中的blobs_是以BlobProto类型存放在layer_param_中
  LayerParameter layer_param_;      //layer的配置参数,protobuf消息类型
  /** The phase: TRAIN or TEST */
  Phase phase_;                     //layer的字段,训练模式(TRAIN)还是测试模式(TEST)
  /** The vector that stores the learnable parameters as a set of blobs. */
  vector<shared_ptr<Blob<Dtype> > > blobs_;   //layer中的可学习参数(如卷积层的权重,偏置参数都是以此类型存放)
  /** Vector indicating whether to compute the diff of each param blob. */
  vector<bool> param_propagate_down_;         //只是每个参数blob是否需要计算梯度

  /** The vector that indicates whether each top blob has a non-zero weight in
   *  the objective function. */
  vector<Dtype> loss_;    //个数与top.size()相等,表示每个top blob在计算loss时的权重,非loss layer中默认都是0

  /** @brief Using the CPU device, compute the layer output. */
  virtual void Forward_cpu(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) = 0;     //使用cpu进行前向计算
  /**
   * @brief Using the GPU device, compute the layer output.
   *        Fall back to Forward_cpu() if unavailable.
   */
  virtual void Forward_gpu(const vector<Blob<Dtype>*>& bottom,
      const vector<Blob<Dtype>*>& top) {      //使用gpu进行前向计算
    // LOG(WARNING) << "Using CPU code as backup.";
    return Forward_cpu(bottom, top);          //子类中未实现则默认调用Forward_cpu()
  }

  /**
   * @brief Using the CPU device, compute the gradients for any parameters and
   *        for the bottom blobs if propagate_down is true.
   */
  virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
      const vector<bool>& propagate_down,
      const vector<Blob<Dtype>*>& bottom) = 0;    //使用cpu进行反向计算
  /**
   * @brief Using the GPU device, compute the gradients for any parameters and
   *        for the bottom blobs if propagate_down is true.
   *        Fall back to Backward_cpu() if unavailable.
   */
  virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,
      const vector<bool>& propagate_down,
      const vector<Blob<Dtype>*>& bottom) {       //使用gpu进行反向计算
    // LOG(WARNING) << "Using CPU code as backup.";
    Backward_cpu(top, propagate_down, bottom);    //未实现则调用Backward_cpu()
  }

  /**
   * Called by the parent Layer's SetUp to check that the number of bottom
   * and top Blobs provided as input match the expected numbers specified by
   * the {ExactNum,Min,Max}{Bottom,Top}Blobs() functions.
   */
  virtual void CheckBlobCounts(const vector<Blob<Dtype>*>& bottom,
                               const vector<Blob<Dtype>*>& top) {   //检查输入输出blob的个数是否符合要求
    if (ExactNumBottomBlobs() >= 0) {
      CHECK_EQ(ExactNumBottomBlobs(), bottom.size())    //指定了输入blob的个数,检查是否相等
          << type() << " Layer takes " << ExactNumBottomBlobs()
          << " bottom blob(s) as input.";
    }
    if (MinBottomBlobs() >= 0) {
      CHECK_LE(MinBottomBlobs(), bottom.size())         //指定了输入blob的最小个数,检查是否小于等于
          << type() << " Layer takes at least " << MinBottomBlobs()
          << " bottom blob(s) as input.";
    }
    if (MaxBottomBlobs() >= 0) {
      CHECK_GE(MaxBottomBlobs(), bottom.size())         //指定了输入blob的最大个数,检查是否大于等于
          << type() << " Layer takes at most " << MaxBottomBlobs()
          << " bottom blob(s) as input.";
    }
    if (ExactNumTopBlobs() >= 0) {
      CHECK_EQ(ExactNumTopBlobs(), top.size())          //指定了输出blob的个数,检查是否相等
          << type() << " Layer produces " << ExactNumTopBlobs()
          << " top blob(s) as output.";
    }
    if (MinTopBlobs() >= 0) {
      CHECK_LE(MinTopBlobs(), top.size())               //指定了输出blob的最小个数,检查是否小于等于
          << type() << " Layer produces at least " << MinTopBlobs()
          << " top blob(s) as output.";
    }
    if (MaxTopBlobs() >= 0) {
      CHECK_GE(MaxTopBlobs(), top.size())               //指定了输出blob的最大个数,检查是否大于等于
          << type() << " Layer produces at most " << MaxTopBlobs()
          << " top blob(s) as output.";
    }
    if (EqualNumBottomTopBlobs()) {
      CHECK_EQ(bottom.size(), top.size())               //指定了输入输出blob的个数必须相等,检查是否相等
          << type() << " Layer produces one top blob as output for each "
          << "bottom blob input.";
    }
  }

  /**
   * Called by SetUp to initialize the weights associated with any top blobs in
   * the loss function. Store non-zero loss weights in the diff blob.
   */
  inline void SetLossWeights(const vector<Blob<Dtype>*>& top) {     //为每个输出blob设置对应的权重
    const int num_loss_weights = layer_param_.loss_weight_size();   //layer_param_中设置的loss weight的个数
    if (num_loss_weights) {
      CHECK_EQ(top.size(), num_loss_weights) << "loss_weight must be "
          "unspecified or specified once per top blob.";            //检查loss权重个数与输出blob个数是否相等
      for (int top_id = 0; top_id < top.size(); ++top_id) {
        const Dtype loss_weight = layer_param_.loss_weight(top_id); //第top_id个输出blob对应的loss权重
        if (loss_weight == Dtype(0)) { continue; }                  //loss权重为0,直接跳过
        this->set_loss(top_id, loss_weight);                        //权重保存在loss_中,loss_[top_id]=loss_weight
        const int count = top[top_id]->count();                     //第top_id个输出blob的数据的个数
        Dtype* loss_multiplier = top[top_id]->mutable_cpu_diff();   //第top_id个输出blob的diff的数据指针
        caffe_set(count, loss_weight, loss_multiplier);   //将输出blob的diff数据设置为对应的权重,loss_multiplier[i]=loss_weight,i=[0,count)
      }
    }
  }

 private:
  DISABLE_COPY_AND_ASSIGN(Layer);
};  // class Layer

// Forward and backward wrappers. You should implement the cpu and
// gpu specific implementations instead, and should not change these
// functions.
template <typename Dtype>
inline Dtype Layer<Dtype>::Forward(const vector<Blob<Dtype>*>& bottom,    //前向计算过程,由输入blob计算输出blob和对应的loss
    const vector<Blob<Dtype>*>& top) {
  Dtype loss = 0;
  Reshape(bottom, top);         //先调整输出blob和内部缓冲区的形状
  switch (Caffe::mode()) {      //当前caffe运行的模式
  case Caffe::CPU:
    Forward_cpu(bottom, top);   //cpu模式下,调用Forward_cpu()执行前向计算
    for (int top_id = 0; top_id < top.size(); ++top_id) {   //每个输出blob数据
      if (!this->loss(top_id)) { continue; }                //loss权重为0,跳过
      const int count = top[top_id]->count();               //blob数据的个数
      const Dtype* data = top[top_id]->cpu_data();          //blob数据的cpu指针
      const Dtype* loss_weights = top[top_id]->cpu_diff();  //blob数据的loss权重
      //loss_weights只有在loss layer中才是非0值.而loss layer中的top blob的data_数据就是误差值,即此处的data存放layer的误差,点乘得到总的loss
      loss += caffe_cpu_dot(count, data, loss_weights);     //data点乘loss_weights,得到loss值
    }
    break;
  case Caffe::GPU:              //caffe运行在gpu模式下
    Forward_gpu(bottom, top);   //调用Forward_gpu()执行前向计算过程
#ifndef CPU_ONLY
    for (int top_id = 0; top_id < top.size(); ++top_id) {   //与上类似,提取输出blob的gpu数据并计算总的loss值
      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;
    }
#endif
    break;
  default:
    LOG(FATAL) << "Unknown caffe mode.";
  }
  return loss;    //返回loss值
}

template <typename Dtype>
inline void Layer<Dtype>::Backward(const vector<Blob<Dtype>*>& top,
    const vector<bool>& propagate_down,
    const vector<Blob<Dtype>*>& bottom) {           //执行反向传播过程
  switch (Caffe::mode()) {      //根据模式调用相应的处理函数
  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.";
  }
}

// Serialize LayerParameter to protocol buffer
//将layer中的参数blob写入到LayerParameter中的BlobProto类型的消息中
template <typename Dtype>
void Layer<Dtype>::ToProto(LayerParameter* param, bool write_diff) {
  param->Clear();         //清空param中的所有消息数据
  param->CopyFrom(layer_param_);    //将layer中的消息layer_param_拷贝至param中
  param->clear_blobs();   //清空param中的BlobProto消息数据
  for (int i = 0; i < blobs_.size(); ++i) {
    //调用blob.cpp中的Blob<double/float>::ToProto()函数,将blob数据以BlobProto消息格式写入param中
    blobs_[i]->ToProto(param->add_blobs(), write_diff);
  }
}

小结

1. Layer类中的成员vector<Dtype> loss_;，loss_中存放的值是输出top blob对应的loss权重，而不是实际的loss值，变量名好像有点误导性。
2. ToProto()函数中先是进行param->clear_blobs()操作，再进行blobs_[i]->ToProto()操作。在创建Layer类的时候，layer中的参数blob是从layer_param_的BlobProto类型的消息中读取进来的（blobs_[i]->FromProto(layer_param_.blobs(i))），此时在Layer类中，LayerParameter中的BlobProto类型的变量与Blob类型的变量的数据内容是一样的。但是随着网络的训练，Blob类型的变量（可学习参数）会不断更新，而LayerParameter中的BlobProto中的数据并不会更新，数据并不一致，所以在ToProto()函数中需要先清除旧的BlobProto数据，将Blob类型数据转化为BlobProto类型存入其中。

Caffe的源码笔者是第一次阅读，一边阅读一边记录，对代码的理解和分析可能会存在错误或遗漏，希望各位读者批评指正，谢谢支持！