本系列参考<深度学习21天实战caffe>这本书所做的笔记,如果错误欢迎指导
前篇
caffe 源码导读(一) 了解protobuf
caffe 源码导读(二) Blob数据结构介绍
caffe 源码导读(三) Blob.hpp头文件解析
d===========================================================================d
caffe源码导读(四)Blob.cpp解析
直接看代码解析
#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; // 用于计算元素总数
shape_.resize(shape.size()); // 成员变量也被重置,调整shape_的维度数和输入的一样
// 如果 shape_data_(blob)的训练数据未分配内存资源,或者指向存储空间小鱼输入shape形状的大小
// 重新分配shape大小的内存空间
if (!shape_data_ || shape_data_->size() < shape.size() * sizeof(int)) {
shape_data_.reset(new SyncedMemory(shape.size() * sizeof(int)));
}
// 类型转换,转换成指向int数据类型指针
int* shape_data = static_cast<int*>(shape_data_->mutable_cpu_data());
// 逐个维度检查对应形状大小是否超出int秒速能力,并累计计算占的空间count_
for (int i = 0; i < shape.size(); ++i) {
CHECK_GE(shape[i], 0); //保证每一个维度的尺度>=0
if (count_ != 0) { // INT_MAX=2^31-1,为了保证count_不溢出
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]; // 形状数据赋值
shape_data[i] = shape[i]; //形状数据指针,shape_data是上面第33行出现的
}
// 如果新的count_大于当前已分配空间容量,重新分配内存空间
// capacity_在头文件中定义的,blob的容积量
if (count_ > capacity_) {
capacity_ = count_;
// shared_ptr,通过reset方式来重新赋值,开辟新的动态内存空间给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()); //创一个vec,再调用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来初始化
}
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
// 调用Reshape之前必须初始化,这些都是构造函数里的
: 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);
}
//获取Blob在gpu存储数据形状的指针,const表示无法通过指针修改数据
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_);//保证data不为空
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); //设置成员变量的值为传入的参数值
}
template <typename Dtype>
const Dtype* Blob<Dtype>::gpu_data() const {
CHECK(data_);
return (const Dtype*)data_->gpu_data(); //只读获取gpu_data的指针
}
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);//设置成员变量的值为传入的参数值 gpu了
}
template <typename Dtype>
const Dtype* Blob<Dtype>::cpu_diff() const {
CHECK(diff_);
return (const Dtype*)diff_->cpu_data();// 只读获取cpu_diff指针
}
template <typename Dtype>
const Dtype* Blob<Dtype>::gpu_diff() const {
CHECK(diff_);
return (const Dtype*)diff_->gpu_data();// 只读获取gpu_diff指针
}
template <typename Dtype>
Dtype* Blob<Dtype>::mutable_cpu_data() { //读写访问cpu_data指针
CHECK(data_);
return static_cast<Dtype*>(data_->mutable_cpu_data());
}
template <typename Dtype>
Dtype* Blob<Dtype>::mutable_gpu_data() { //读写访问gpu_data指针
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) { // 共享另一个Blob的data指针
CHECK_EQ(count_, other.count());
data_ = other.data();
}
template <typename Dtype>
void Blob<Dtype>::ShareDiff(const Blob& other) { // 共享另一个Blob的diff指针
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用于网络参数blob更新,只有float个double实现,并没有int和unsigned实现
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在哪里,就在哪里更新
switch (data_->head()) {
case SyncedMemory::HEAD_AT_CPU: // data位于cpu端
// perform computation on CPU
// 执行cpu上的计算,data_[i] = data_[i] - diff_[i], i=0,1,...,count-1
// 在util/math_functions.cpp中,调用了cblas_saxpy函数
//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); }实现 0-n中,-1*diff + data_
// 有一篇博客总结了 caffe调用的cblas函数
// 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位于gpu端,或者cpu/gpu同步
case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
// perform computation on GPU
// GPU上的计算 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.";
}
}
// 计算L1范数,但是这边好像并不能使用
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为空,返回0
switch (data_->head()) { // 选择data_的指针头是在cpu上还是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:计算 vector x 的所有element的绝对值之和。
}
**/
case SyncedMemory::HEAD_AT_CPU:
return caffe_cpu_asum(count_, cpu_data());// cpu上L1范数计算,即绝对值之和
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()); // 计算diff的L1范数
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;
}
// 计算data_的L2范数
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);//执行cpu上的dot计算
break;
case SyncedMemory::HEAD_AT_GPU:
case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
data = gpu_data();
caffe_gpu_dot(count_, data, data, &sumsq);
#else
NO_GPU;
#endif
break;
case SyncedMemory::UNINITIALIZED:
return 0;
default:
LOG(FATAL) << "Unknown SyncedMemory head state: " << data_->head();
}
return sumsq;
}
template <> unsigned int Blob<unsigned int>::sumsq_diff() const {
NOT_IMPLEMENTED;
return 0;
}
template <> int Blob<int>::sumsq_diff() const {
NOT_IMPLEMENTED;
return 0;
}
// 计算diff上的L2范数
template <typename Dtype>
Dtype Blob<Dtype>::sumsq_diff() const {
Dtype sumsq;
const Dtype* diff;
if (!diff_) { return 0; }
switch (diff_->head()) {
case SyncedMemory::HEAD_AT_CPU:
diff = cpu_diff();
sumsq = caffe_cpu_dot(count_, diff, diff);
break;
case SyncedMemory::HEAD_AT_GPU:
case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
diff = gpu_diff();
caffe_gpu_dot(count_, diff, diff, &sumsq);
break;
#else
NO_GPU;
#endif
case SyncedMemory::UNINITIALIZED:
return 0;
default:
LOG(FATAL) << "Unknown SyncedMemory head state: " << data_->head();
}
return sumsq;
}
template <> void Blob<unsigned int>::scale_data(unsigned int scale_factor) {
NOT_IMPLEMENTED;
}
template <> void Blob<int>::scale_data(int scale_factor) {
NOT_IMPLEMENTED;
}
// 对data_进行幅度缩放
template <typename Dtype>
void Blob<Dtype>::scale_data(Dtype scale_factor) {
Dtype* data;
if (!data_) { return; }
switch (data_->head()) {
case SyncedMemory::HEAD_AT_CPU:
data = mutable_cpu_data(); //可读写指针
caffe_scal(count_, scale_factor, data); //data[i] = data[i]*scale_factor
return;
case SyncedMemory::HEAD_AT_GPU:
case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
data = mutable_gpu_data();
caffe_gpu_scal(count_, scale_factor, data);
return;
#else
NO_GPU;
#endif
case SyncedMemory::UNINITIALIZED:
return;
default:
LOG(FATAL) << "Unknown SyncedMemory head state: " << data_->head();
}
}
template <> void Blob<unsigned int>::scale_diff(unsigned int scale_factor) {
NOT_IMPLEMENTED;
}
template <> void Blob<int>::scale_diff(int scale_factor) {
NOT_IMPLEMENTED;
}
// 同上,对diff梯度幅度缩放
template <typename Dtype>
void Blob<Dtype>::scale_diff(Dtype scale_factor) {
Dtype* diff;
if (!diff_) { return; }
switch (diff_->head()) {
case SyncedMemory::HEAD_AT_CPU:
diff = mutable_cpu_diff();
caffe_scal(count_, scale_factor, diff);
return;
case SyncedMemory::HEAD_AT_GPU:
case SyncedMemory::SYNCED:
#ifndef CPU_ONLY
diff = mutable_gpu_diff();
caffe_gpu_scal(count_, scale_factor, diff);
return;
#else
NO_GPU;
#endif
case SyncedMemory::UNINITIALIZED:
return;
default:
LOG(FATAL) << "Unknown SyncedMemory head state: " << diff_->head();
}
}
// 判断形状是否相同,依次验证NCWH
template <typename Dtype>
bool Blob<Dtype>::ShapeEquals(const BlobProto& other) {
if (other.has_num() || other.has_channels() ||
other.has_height() || other.has_width()) {
// Using deprecated 4D Blob dimensions --
// 使用以及被抛弃的blob维度 版本
// shape is (num, channels, height, width).
// Note: we do not use the normal Blob::num(), Blob::channels(), etc.
// methods as these index from the beginning of the blob shape, where legacy
// parameter blobs were indexed from the end of the blob shape (e.g., bias
// Blob shape (1 x 1 x 1 x N), IP layer weight Blob shape (1 x 1 x M x N)).
// 输入维度使用过时的维度信息,则需要转换为新的vector
return shape_.size() <= 4 &&
LegacyShape(-4) == other.num() &&
LegacyShape(-3) == other.channels() &&
LegacyShape(-2) == other.height() &&
LegacyShape(-1) == other.width();
}
// 直接对比
vector<int> other_shape(other.shape().dim_size());
for (int i = 0; i < other.shape().dim_size(); ++i) {
other_shape[i] = other.shape().dim(i);
}
return shape_ == other_shape;
}
// 从source blob复制数据,必要时变维
template <typename Dtype>
void Blob<Dtype>::CopyFrom(const Blob& source, bool copy_diff, bool reshape) {
if (source.count() != count_ || source.shape() != shape_) {
if (reshape) {
ReshapeLike(source); //要求变维,就得变维
} else { // 不变维,两个blob的形状又不同,拷贝不了啊
LOG(FATAL) << "Trying to copy blobs of different sizes.";
}
}
switch (Caffe::mode()) {
case Caffe::GPU: 如果使用cpu
if (copy_diff) {
caffe_copy(count_, source.gpu_diff(),
static_cast<Dtype*>(diff_->mutable_gpu_data()));
} else {
caffe_copy(count_, source.gpu_data(),
static_cast<Dtype*>(data_->mutable_gpu_data()));
}
break;
case Caffe::CPU: //如果使用cpu
if (copy_diff) {
caffe_copy(count_, source.cpu_diff(),
static_cast<Dtype*>(diff_->mutable_cpu_data()));
} else {
caffe_copy(count_, source.cpu_data(),
static_cast<Dtype*>(data_->mutable_cpu_data()));
}
break;
default:
LOG(FATAL) << "Unknown caffe mode.";
}
}
// 从BlobProto中加载一个Blob, 适用于从磁盘载入之前导出的Blob
template <typename Dtype>
void Blob<Dtype>::FromProto(const BlobProto& proto, bool reshape) {
if (reshape) { // 从BlobProto中获取所需各个维度信息
vector<int> shape;
if (proto.has_num() || proto.has_channels() ||
proto.has_height() || proto.has_width()) {
// Using deprecated 4D Blob dimensions --
// shape is (num, channels, height, width).
// 这边已经过时的维度信息
shape.resize(4);
shape[0] = proto.num();
shape[1] = proto.channels();
shape[2] = proto.height();
shape[3] = proto.width();
} else {
shape.resize(proto.shape().dim_size());
for (int i = 0; i < proto.shape().dim_size(); ++i) {
shape[i] = proto.shape().dim(i);
}
}
Reshape(shape); // 变维,按照bLOB进行变维
} else {
CHECK(ShapeEquals(proto)) << "shape mismatch (reshape not set)";
}
// copy data 复制数据
Dtype* data_vec = mutable_cpu_data();
if (proto.double_data_size() > 0) { //如果之前保存的是double的data,加载double的data
CHECK_EQ(count_, proto.double_data_size());
for (int i = 0; i < count_; ++i) {
data_vec[i] = proto.double_data(i);//加载double的data
}
} else {
CHECK_EQ(count_, proto.data_size());
for (int i = 0; i < count_; ++i) {
data_vec[i] = proto.data(i);// 否则加载 float的data
}
}
// diff 同理
if (proto.double_diff_size() > 0) {
CHECK_EQ(count_, proto.double_diff_size());
Dtype* diff_vec = mutable_cpu_diff();
for (int i = 0; i < count_; ++i) {
diff_vec[i] = proto.double_diff(i);
}
} else if (proto.diff_size() > 0) {
CHECK_EQ(count_, proto.diff_size());
Dtype* diff_vec = mutable_cpu_diff();
for (int i = 0; i < count_; ++i) {
diff_vec[i] = proto.diff(i);
}
}
}
// 将Blob中的data导出到BlobProto结构体,便于存储
template <>
void Blob<double>::ToProto(BlobProto* proto, bool write_diff) const {
proto->clear_shape(); // 重置维度,保证维度相同
for (int i = 0; i < shape_.size(); ++i) {
proto->mutable_shape()->add_dim(shape_[i]);
}
proto->clear_double_data(); //清楚数据
proto->clear_double_diff();
const double* data_vec = cpu_data();
for (int i = 0; i < count_; ++i) {
proto->add_double_data(data_vec[i]);
}
if (write_diff) {
const double* diff_vec = cpu_diff();
for (int i = 0; i < count_; ++i) {
proto->add_double_diff(diff_vec[i]);
}
}
}
// 同上,float
template <>
void Blob<float>::ToProto(BlobProto* proto, bool write_diff) const {
proto->clear_shape();
for (int i = 0; i < shape_.size(); ++i) {
proto->mutable_shape()->add_dim(shape_[i]);
}
proto->clear_data();
proto->clear_diff();
const float* data_vec = cpu_data(); // 将data导出到proto
for (int i = 0; i < count_; ++i) {
proto->add_data(data_vec[i]);
}
if (write_diff) { //如果需要 写入diff
const float* diff_vec = cpu_diff();
for (int i = 0; i < count_; ++i) {
proto->add_diff(diff_vec[i]);
}
}
}
INSTANTIATE_CLASS(Blob);
template class Blob<int>;
template class Blob<unsigned int>;
} // namespace caffe
参考文献
<深度学习caffe实战21天>
caffe源码 Blob.cpp 详解
Caffe源码(一):math_functions 分析
caffe源码解析 — blob.cpp