本系列參考<深度學習21天實戰caffe>這本書所做的筆記,如果錯誤歡迎指導

前篇

caffe 原始碼導讀(一) 瞭解protobuf
caffe 原始碼導讀(二) Blob資料結構介紹
caffe 原始碼導讀(三) Blob.hpp標頭檔案解析
d===========================================================================d

caffe原始碼導讀(四)Blob.cpp解析

直接看程式碼解析

#include <climits>
#include <vector>
 


#include "caffe/blob.hpp"
#include "caffe/common.hpp"
#include "caffe/syncedmem.hpp"
#include "caffe/util/math_functions.hpp"

namespace caffe {

template <typename Dtype>
void Blob<Dtype>::Reshape(const int num, const int channels, const int height,
    const int width) {
  vector<int>
 
 shape(4);
  shape[0] = num;
  shape[1] = channels;
  shape[2] = height;
  shape[3] = width;
  Reshape(shape);
}

template <typename Dtype>
void Blob<Dtype>::Reshape(const vector<int>& shape) {
  CHECK_LE(shape.size(), kMaxBlobAxes);
  count_ = 1; // 用於計算元素總數
  shape_.resize(shape.
 
size()); // 成員變數也被重置,調整shape_的維度數和輸入的一樣
  // 如果 shape_data_(blob)的訓練資料未分配記憶體資源,或者指向儲存空間小魚輸入shape形狀的大小
  // 重新分配shape大小的記憶體空間
  if (!shape_data_ || shape_data_->size() < shape.size() * sizeof(int)) {
    shape_data_.reset(new SyncedMemory(shape.size() * sizeof(int)));
  }
  // 型別轉換,轉換成指向int資料型別指標
  int* shape_data = static_cast<int*>(shape_data_->mutable_cpu_data());
  // 逐個維度檢查對應形狀大小是否超出int秒速能力,並累計計算佔的空間count_
  for (int i = 0; i < shape.size(); ++i) {
    CHECK_GE(shape[i], 0); //保證每一個維度的尺度>=0
    if (count_ != 0) { // INT_MAX=2^31-1,為了保證count_不溢位
      CHECK_LE(shape[i], INT_MAX / count_) << "blob size exceeds INT_MAX";
    }
    count_ *= shape[i];// count累乘, 先乘n,再c,再h,再w
    shape_[i] = shape[i]; // 形狀資料賦值
    shape_data[i] = shape[i]; //形狀資料指標,shape_data是上面第33行出現的
  }
    // 如果新的count_大於當前已分配空間容量,重新分配記憶體空間
    // capacity_在標頭檔案中定義的,blob的容積量
  if (count_ > capacity_) {
    capacity_ = count_;
    // shared_ptr,通過reset方式來重新賦值,開闢新的動態記憶體空間給blob
    data_.reset(new SyncedMemory(capacity_ * sizeof(Dtype)));
    diff_.reset(new SyncedMemory(capacity_ * sizeof(Dtype)));
  }
}

template <typename Dtype>
void Blob<Dtype>::Reshape(const BlobShape& shape) {
  CHECK_LE(shape.dim_size(), kMaxBlobAxes);
  vector<int> shape_vec(shape.dim_size()); //創一個vec,再呼叫Reshape
  for (int i = 0; i < shape.dim_size(); ++i) {
    shape_vec[i] = shape.dim(i);
  }
  Reshape(shape_vec);
}

template <typename Dtype>
void Blob<Dtype>::ReshapeLike(const Blob<Dtype>& other) {
  Reshape(other.shape()); //其他Blob的shape來初始化
}

template <typename Dtype>
Blob<Dtype>::Blob(const int num, const int channels, const int height,
    const int width)
  // capacity_ must be initialized before calling Reshape
  // 呼叫Reshape之前必須初始化,這些都是建構函式裡的
  : capacity_(0) {
  Reshape(num, channels, height, width);
}

template <typename Dtype>
Blob<Dtype>::Blob(const vector<int>& shape)
  // capacity_ must be initialized before calling Reshape
  : capacity_(0) {
  Reshape(shape);
}

//獲取Blob在gpu儲存資料形狀的指標,const表示無法通過指標修改資料
template <typename Dtype>
const int* Blob<Dtype>::gpu_shape() const {
  CHECK(shape_data_);
  return (const int*)shape_data_->gpu_data();
}

template <typename Dtype>
const Dtype* Blob<Dtype>::cpu_data() const {
  CHECK(data_);//保證data不為空
  return (const Dtype*)data_->cpu_data();
}

template <typename Dtype>
void Blob<Dtype>::set_cpu_data(Dtype* data) {
  CHECK(data);
  // Make sure CPU and GPU sizes remain equal
  size_t size = count_ * sizeof(Dtype);
  if (data_->size() != size) {
    data_.reset(new SyncedMemory(size));
    diff_.reset(new SyncedMemory(size));
  }
  data_->set_cpu_data(data); //設定成員變數的值為傳入的引數值
}

template <typename Dtype>
const Dtype* Blob<Dtype>::gpu_data() const {
  CHECK(data_);
  return (const Dtype*)data_->gpu_data(); //只讀獲取gpu_data的指標
}

template <typename Dtype>
void Blob<Dtype>::set_gpu_data(Dtype* data) {
  CHECK(data);
  // Make sure CPU and GPU sizes remain equal
  size_t size = count_ * sizeof(Dtype);
  if (data_->size() != size) {
    data_.reset(new SyncedMemory(size));
    diff_.reset(new SyncedMemory(size));
  }
  data_->set_gpu_data(data);//設定成員變數的值為傳入的引數值 gpu了
}

template <typename Dtype>
const Dtype* Blob<Dtype>::cpu_diff() const {
  CHECK(diff_);
  return (const Dtype*)diff_->cpu_data();// 只讀獲取cpu_diff指標
}

template <typename Dtype>
const Dtype* Blob<Dtype>::gpu_diff() const {
  CHECK(diff_);
  return (const Dtype*)diff_->gpu_data();// 只讀獲取gpu_diff指標
}

template <typename Dtype>
Dtype* Blob<Dtype>::mutable_cpu_data() { //讀寫訪問cpu_data指標
  CHECK(data_);
  return static_cast<Dtype*>(data_->mutable_cpu_data());
}

template <typename Dtype>
Dtype* Blob<Dtype>::mutable_gpu_data() { //讀寫訪問gpu_data指標
  CHECK(data_);
  return static_cast<Dtype*>(data_->mutable_gpu_data());
}

template <typename Dtype>
Dtype* Blob<Dtype>::mutable_cpu_diff() {
  CHECK(diff_);
  return static_cast<Dtype*>(diff_->mutable_cpu_data());
}

template <typename Dtype>
Dtype* Blob<Dtype>::mutable_gpu_diff() {
  CHECK(diff_);
  return static_cast<Dtype*>(diff_->mutable_gpu_data());
}

template <typename Dtype>
void Blob<Dtype>::ShareData(const Blob& other) { // 共享另一個Blob的data指標
  CHECK_EQ(count_, other.count());
  data_ = other.data();
}

template <typename Dtype>
void Blob<Dtype>::ShareDiff(const Blob& other) { // 共享另一個Blob的diff指標
  CHECK_EQ(count_, other.count());
  diff_ = other.diff();
}

// The "update" method is used for parameter blobs in a Net, which are stored
// as Blob<float> or Blob<double> -- hence we do not define it for
// Blob<int> or Blob<unsigned int>.
// update用於網路引數blob更新,只有float個double實現,並沒有int和unsigned實現
template <> void Blob<unsigned int>::Update() { NOT_IMPLEMENTED; }
template <> void Blob<int>::Update() { NOT_IMPLEMENTED; }

template <typename Dtype>
void Blob<Dtype>::Update() {
  // We will perform update based on where the data is located.
  // data在哪裡,就在哪裡更新
  switch (data_->head()) {
  case SyncedMemory::HEAD_AT_CPU: // data位於cpu端
    // perform computation on CPU
    // 執行cpu上的計算,data_[i] = data_[i] - diff_[i], i=0,1,...,count-1
    // 在util/math_functions.cpp中,呼叫了cblas_saxpy函式
    //template <typename Dtype>
  //    void caffe_axpy(const int N, const Dtype alpha, const Dtype* X, Dtype* Y);
  // void caffe_axpy<float>(const int N, const float alpha, const float* X,float* Y) {
  // cblas_saxpy(N, alpha, X, 1, Y, 1); }實現 0-n中,-1*diff + data_
  // 有一篇部落格總結了 caffe呼叫的cblas函式
  // https://www.cnblogs.com/jianyingzhou/p/4444728.html
    caffe_axpy<Dtype>(count_, Dtype(-1),
        static_cast<const Dtype*>(diff_->cpu_data()),
        static_cast<Dtype*>(data_->mutable_cpu_data()));
    break;
  case SyncedMemory::HEAD_AT_GPU: //data位於gpu端,或者cpu/gpu同步
  case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
    // perform computation on GPU
    // GPU上的計算 data_[i] = data_[i] - diff_[i]
    caffe_gpu_axpy<Dtype>(count_, Dtype(-1),
        static_cast<const Dtype*>(diff_->gpu_data()),
        static_cast<Dtype*>(data_->mutable_gpu_data()));
#else
    NO_GPU;
#endif
    break;
  default:
    LOG(FATAL) << "Syncedmem not initialized.";
  }
}

// 計算L1範數,但是這邊好像並不能使用
template <> unsigned int Blob<unsigned int>::asum_data() const {
  NOT_IMPLEMENTED;
  return 0;
}

template <> int Blob<int>::asum_data() const {
  NOT_IMPLEMENTED;
  return 0;
}

template <typename Dtype>
Dtype Blob<Dtype>::asum_data() const {
  if (!data_) { return 0; } // 如果data為空,返回0
  switch (data_->head()) { // 選擇data_的指標頭是在cpu上還是gpu上
  /** Returns the sum of the absolute values of the elements of vector x
  template <typename Dtype>
  Dtype caffe_cpu_asum(const int n, const Dtype* x);
   float caffe_cpu_asum<float>(const int n, const float* x) {
  return cblas_sasum(n, x, 1);
   功能：計算 vector x 的所有element的絕對值之和。
}
   **/
  case SyncedMemory::HEAD_AT_CPU:
    return caffe_cpu_asum(count_, cpu_data());// cpu上L1範數計算,即絕對值之和
  case SyncedMemory::HEAD_AT_GPU:
  case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
  {
    Dtype asum;
    caffe_gpu_asum(count_, gpu_data(), &asum);
    return asum;
  }
#else
    NO_GPU;
#endif
  case SyncedMemory::UNINITIALIZED:
    return 0;
  default:
    LOG(FATAL) << "Unknown SyncedMemory head state: " << data_->head();
  }
  return 0;
}

template <> unsigned int Blob<unsigned int>::asum_diff() const {
  NOT_IMPLEMENTED;
  return 0;
}

template <> int Blob<int>::asum_diff() const {
  NOT_IMPLEMENTED;
  return 0;
}

template <typename Dtype>
Dtype Blob<Dtype>::asum_diff() const {
  if (!diff_) { return 0; }
  switch (diff_->head()) {
  case SyncedMemory::HEAD_AT_CPU:
    return caffe_cpu_asum(count_, cpu_diff()); // 計算diff的L1範數
  case SyncedMemory::HEAD_AT_GPU:
  case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
  {
    Dtype asum;
    caffe_gpu_asum(count_, gpu_diff(), &asum);
    return asum;
  }
#else
    NO_GPU;
#endif
  case SyncedMemory::UNINITIALIZED:
    return 0;
  default:
    LOG(FATAL) << "Unknown SyncedMemory head state: " << diff_->head();
  }
  return 0;
}

template <> unsigned int Blob<unsigned int>::sumsq_data() const {
  NOT_IMPLEMENTED;
  return 0;
}

template <> int Blob<int>::sumsq_data() const {
  NOT_IMPLEMENTED;
  return 0;
}

// 計算data_的L2範數
template <typename Dtype>
Dtype Blob<Dtype>::sumsq_data() const {
  Dtype sumsq;
  const Dtype* data;
  if (!data_) { return 0; }
  switch (data_->head()) {
  case SyncedMemory::HEAD_AT_CPU:
    data = cpu_data();
    sumsq = caffe_cpu_dot(count_, data, data);//執行cpu上的dot計算
    break;
  case SyncedMemory::HEAD_AT_GPU:
  case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
    data = gpu_data();
    caffe_gpu_dot(count_, data, data, &sumsq)<