Caffe源代碼(caffe version commit: 09868ac , date: 2015.08.15)中有一些重要的頭文件，這裏介紹下include/caffe/solver.hpp文件的內容：

1. include文件：

<caffe/solver.hpp>：此文件的介紹能夠參考： http://blog.csdn.net/fengbingchun/article/details/62423060

2. 模板類Solver：虛基類

3. 模板類WorkerSolver：繼承父類Solver，用於多GPU訓練時僅計算梯度

4. 模板類SGDSolver：繼承父類Solver

5. 模板類NesterovSolver：繼承SGDSolver

6. 模板類AdaGradSolver：繼承SGDSolver

7. 模板類RMSPropSolver：繼承SGDSolver

8. 模板類AdaDeltaSolver：繼承SGDSolver

9. 模板類AdamSolver：繼承SGDSolver

10. 函數GetSolver：new solver對象

Solver通過協調Net的前向判斷計算和反向梯度計算(forward inference and backward gradients)，來對參數進行更新。從而達到降低loss的目的。Caffe模型的學習被分為兩個部分：由Solver進行優化、更新參數。由Net計算出loss和gradient。

solver.prototxt是一個配置文件用來告知Caffe如何對網絡進行訓練。

有了Net就能夠進行神經網絡的前後向傳播計算了。可是還缺少神經網絡的訓練和預測功能，Solver類進一步封裝了訓練和預測相關的一些功能。Solver定義了針對Net網絡模型的求解方法，記錄神經網絡的訓練過程，保存神經網絡模型參數，中斷並恢復網絡的訓練過程。自己定義Solver能夠實現不同的神經網絡求解方式。

Caffe支持的solvers包含：

(1)、Stochastic Gradient Descent(type: “SGD”)即隨機梯度下降：利用負梯度和上一次權重的更新值的線性組合來更新權重。學習率(learning rate)是負梯度的權重。

動量是上一次更新值的權重。一般將學習速率初始化為0.01。然後在訓練(training)中當loss達到穩定時，將學習速率除以一個常數(比如10)，將這個過程重復多次。

對於動量一般設置為0.9，動量使weight得更新更為平緩，使學習過程更為穩定、高速。

(2)、AdaDelta(type:“AdaDelta”)：是一種”魯棒的學習率方法”，同SGD一樣是一種基於梯度的優化方法。

(3)、Adaptive Gradient(type: “AdaGrad”)即自適應梯度下降，與隨機梯度下降一樣是基於梯度的優化方法。

(4)、Adam(type:“Adam”)：也是一種基於梯度的優化方法。

它包含一對自適應時刻預計變量，能夠看做是AdaGrad的一種泛化形式。

(5)、Nesterov’s Accelerated Gradient(type: “Nesterov”)：Nesterov提出的加速梯度下降(Nesterov’s accelerated gradient)是凸優化的一種最優算法，其收斂速度能夠達到O(1/t^2)，而不是O(1/t)。

雖然在使用Caffe訓練深度神經網絡時非常難滿足O(1/t^2)收斂條件。但實際中NAG對於某些特定結構的深度學習模型仍是一個非常有效的方法。

(6)、RMSprop(type:“RMSProp”)：是一種基於梯度的優化方法(同SGD相似)。

Solver：

(1)、用於優化過程的記錄、創建訓練網絡(用於學習)和測試網絡(用於評估)；

(2)、通過forward和backward過程來叠代地優化和更新參數；

(3)、周期性地用測試網絡評估模型性能；

(4)、在優化過程中記錄模型和solver狀態的快照(snapshot)。

每一次叠代過程中：

(1)、調用Net的前向過程計算出輸出和loss。

(2)、調用Net的反向過程計算出梯度(loss對每層的權重w和偏置b求導)。

(3)、依據以下所講的Solver方法。利用梯度更新參數；

(4)、依據學習率(learning rate)。歷史數據和求解方法更新solver的狀態。使權重從初始化狀態逐步更新到終於的學習到的狀態。

Solvers的運行模式有CPU/GPU兩種模式。

Solver方法：用於最小化損失(loss)值。

給定一個數據集D，優化的目標是D中全部數據損失的均值，即平均損失。取得最小值。

註：以上關於Solver內容的介紹主要摘自由CaffeCN社區翻譯的《Caffe官方教程中譯本》。

<caffe/solver.hpp>文件的具體介紹例如以下：

#ifndef CAFFE_OPTIMIZATION_SOLVER_HPP_
#define CAFFE_OPTIMIZATION_SOLVER_HPP_

#include <string>
#include <vector>

#include "caffe/net.hpp"

namespace caffe {

/**
 * @brief An interface for classes that perform optimization on Net%s.
 *
 * Requires implementation of ApplyUpdate to compute a parameter update
 * given the current state of the Net parameters.
 */
template <typename Dtype>
class Solver { // Solver模板類，虛基類
 public:
// 顯示構造函數, 內部會調用Init函數
  explicit Solver(const SolverParameter& param, const Solver* root_solver = NULL);
  explicit Solver(const string& param_file, const Solver* root_solver = NULL);
// 成員變量賦值，包含param_、iter_、current_step_,並調用InitTrainNet和InitTestNets函數
  void Init(const SolverParameter& param);
// 為成員變量net_賦值
  void InitTrainNet();
// 為成員變量test_nets_賦值
  void InitTestNets();
  // The main entry of the solver function. In default, iter will be zero. Pass
  // in a non-zero iter number to resume training for a pre-trained net.
// 依次調用函數Restore、Step、Snapshot，然後運行net_的前向傳播函數ForwardPrefilled，最後調用TestAll函數
  virtual void Solve(const char* resume_file = NULL);
  inline void Solve(const string resume_file) { Solve(resume_file.c_str()); }
// 重復運行net前向傳播反向傳播計算,期間會調用函數TestAll、ApplyUpdate、Snapshot及類Callback兩個成員函數
  void Step(int iters);
  // The Restore method simply dispatches to one of the
  // RestoreSolverStateFrom___ protected methods. You should implement these
  // methods to restore the state from the appropriate snapshot type.
// 載入已有的模型
  void Restore(const char* resume_file);
// 虛析構函數
  virtual ~Solver() {}
 // 獲得slover parameter
  inline const SolverParameter& param() const { return param_; }
// 獲得train Net
  inline shared_ptr<Net<Dtype> > net() { return net_; }
// 獲得test Net
  inline const vector<shared_ptr<Net<Dtype> > >& test_nets() {
    return test_nets_;
  }
// 獲得當前的叠代數
  int iter() { return iter_; }
  // Invoked at specific points during an iteration
// 內部Callback類，僅在多卡GPU模式下使用
  class Callback {
   protected:
    virtual void on_start() = 0;
    virtual void on_gradients_ready() = 0;

    template <typename T>
    friend class Solver;
  };
// 獲得Callback
  const vector<Callback*>& callbacks() const { return callbacks_; }
// 加入一個Callback
  void add_callback(Callback* value) { callbacks_.push_back(value); }

 protected:
  // Make and apply the update value for the current iteration.
// 更新net的權值和偏置
  virtual void ApplyUpdate() = 0;
  // The Solver::Snapshot function implements the basic snapshotting utility
  // that stores the learned net. You should implement the SnapshotSolverState()
  // function that produces a SolverState protocol buffer that needs to be
  // written to disk together with the learned net.
// 快照，內部會調用SnapshotToBinaryProto或SnapshotToHDF5、SnapshotSolverState函數
  void Snapshot();
// 獲取快照文件名稱
  string SnapshotFilename(const string extension);
// 寫proto到.caffemodel
  string SnapshotToBinaryProto();
// 寫proto到HDF5文件
  string SnapshotToHDF5();
  // The test routine
// 內部會循環調用Test函數
  void TestAll();
// 運行測試網絡。net前向傳播
  void Test(const int test_net_id = 0);
// 存儲snapshot solver state
  virtual void SnapshotSolverState(const string& model_filename) = 0;
// 讀HDF5文件到solver state
  virtual void RestoreSolverStateFromHDF5(const string& state_file) = 0;
// 讀二進制文件.solverstate到solver state
  virtual void RestoreSolverStateFromBinaryProto(const string& state_file) = 0;
// dummy function，僅僅有聲明沒有實現
  void DisplayOutputBlobs(const int net_id);

// Caffe中類的成員變量名都帶有後綴"_"。這樣就easy區分暫時變量和類成員變量
  SolverParameter param_; // solver parameter
  int iter_; // 當前的叠代數
  int current_step_; // 
  shared_ptr<Net<Dtype> > net_; // train net
  vector<shared_ptr<Net<Dtype> > > test_nets_; // test net
  vector<Callback*> callbacks_; // Callback

  // The root solver that holds root nets (actually containing shared layers)
  // in data parallelism
  const Solver* const root_solver_;

// 禁止使用Solver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(Solver);
};

/**
 * @brief Solver that only computes gradients, used as worker
 *        for multi-GPU training.
 */
template <typename Dtype>
class WorkerSolver : public Solver<Dtype> { // 模板類WorkerSolver。繼承父類Solver
 public:
// 顯示構造函數
  explicit WorkerSolver(const SolverParameter& param, const Solver<Dtype>* root_solver = NULL)
      : Solver<Dtype>(param, root_solver) {}

 protected:
  void ApplyUpdate() {}
  void SnapshotSolverState(const string& model_filename) {
    LOG(FATAL) << "Should not be called on worker solver.";
  }
  void RestoreSolverStateFromBinaryProto(const string& state_file) {
    LOG(FATAL) << "Should not be called on worker solver.";
  }
  void RestoreSolverStateFromHDF5(const string& state_file) {
    LOG(FATAL) << "Should not be called on worker solver.";
  }
};

/**
 * @brief Optimizes the parameters of a Net using
 *        stochastic gradient descent (SGD) with momentum.
 */
template <typename Dtype>
class SGDSolver : public Solver<Dtype> { // 模板類SGDSolver,繼承父類Solver
 public:
// 顯示構造函數，調用PreSolve函數
  explicit SGDSolver(const SolverParameter& param) : Solver<Dtype>(param) { PreSolve(); }
  explicit SGDSolver(const string& param_file) : Solver<Dtype>(param_file) { PreSolve(); }
// 獲取history數據
  const vector<shared_ptr<Blob<Dtype> > >& history() { return history_; }

 protected:
// 成員變量history_, update_, temp_初始化
  void PreSolve();
// 獲取學習率
  Dtype GetLearningRate();
// 內部會調用ClipGradients、Normalize、Regularize、ComputeUpdateValue，更新net權值和偏置
  virtual void ApplyUpdate();
// 調用caffe_scal函數
  virtual void Normalize(int param_id);
// 調用caffe_axpy函數
  virtual void Regularize(int param_id);
// 計算並更新對應Blob值，調用caffe_cpu_axpby和caffe_copy函數
  virtual void ComputeUpdateValue(int param_id, Dtype rate);
// clip parameter gradients to that L2 norm，假設梯度值過大，就會對梯度做一個修剪。
// 對全部的參數乘以一個縮放因子，使得全部參數的平方和不超過參數中設定的梯度總值
  virtual void ClipGradients();
// 存儲snapshot solver state，內部會掉用SnapshotSolverStateToBinaryProto或SnapshotSolverStateToHDF5函數
  virtual void SnapshotSolverState(const string& model_filename);
// 寫solver state到二進制文件.solverstate
  virtual void SnapshotSolverStateToBinaryProto(const string& model_filename);
// 寫solver state到HDF5
  virtual void SnapshotSolverStateToHDF5(const string& model_filename);
  // 讀HDF5文件到solver state
  virtual void RestoreSolverStateFromHDF5(const string& state_file);
  // 讀二進制文件.solverstate到solver state
  virtual void RestoreSolverStateFromBinaryProto(const string& state_file);
  // history maintains the historical momentum data.
  // update maintains update related data and is not needed in snapshots.
  // temp maintains other information that might be needed in computation
  //   of gradients/updates and is not needed in snapshots
// Caffe中類的成員變量名都帶有後綴"_"，這樣就easy區分暫時變量和類成員變量
  vector<shared_ptr<Blob<Dtype> > > history_, update_, temp_;

// 禁止使用SGDSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(SGDSolver);
};

template <typename Dtype>
class NesterovSolver : public SGDSolver<Dtype> { // 模板類NesterovSolver，繼承SGDSolver
 public:
// 顯示構造函數
  explicit NesterovSolver(const SolverParameter& param) : SGDSolver<Dtype>(param) {}
  explicit NesterovSolver(const string& param_file) : SGDSolver<Dtype>(param_file) {}

 protected:
// 計算並更新對應Blob值,調用caffe_cpu_axpby和caffe_copy函數
  virtual void ComputeUpdateValue(int param_id, Dtype rate);

// 禁止使用NesterovSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(NesterovSolver);
};

template <typename Dtype>
class AdaGradSolver : public SGDSolver<Dtype> { // 模板類AdaGradSolver,繼承SGDSolver
 public:
// 顯示構造函數，調用constuctor_sanity_check函數
  explicit AdaGradSolver(const SolverParameter& param) : SGDSolver<Dtype>(param) { constructor_sanity_check(); }
  explicit AdaGradSolver(const string& param_file) : SGDSolver<Dtype>(param_file) { constructor_sanity_check(); }

 protected:
// 計算並更新對應Blob值
  virtual void ComputeUpdateValue(int param_id, Dtype rate);
  void constructor_sanity_check() {
    CHECK_EQ(0, this->param_.momentum())
        << "Momentum cannot be used with AdaGrad.";
  }

// 禁止使用AdaGradSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(AdaGradSolver);
};


template <typename Dtype>
class RMSPropSolver : public SGDSolver<Dtype> { // 模板類RMSPropSolver,繼承SGDSolver
 public:
// 顯示構造函數。調用constructor_sanity_check函數
  explicit RMSPropSolver(const SolverParameter& param) : SGDSolver<Dtype>(param) { constructor_sanity_check(); }
  explicit RMSPropSolver(const string& param_file) : SGDSolver<Dtype>(param_file) { constructor_sanity_check(); }

 protected:
// 計算並更新對應Blob值
  virtual void ComputeUpdateValue(int param_id, Dtype rate);
  void constructor_sanity_check() {
    CHECK_EQ(0, this->param_.momentum())
        << "Momentum cannot be used with RMSProp.";
    CHECK_GE(this->param_.rms_decay(), 0)
        << "rms_decay should lie between 0 and 1.";
    CHECK_LT(this->param_.rms_decay(), 1)
        << "rms_decay should lie between 0 and 1.";
  }

// 禁止使用RMSPropSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(RMSPropSolver);
};

template <typename Dtype>
class AdaDeltaSolver : public SGDSolver<Dtype> { // 模板類AdaDeltaSolver。繼承SGDSolver
 public:
// 顯示構造函數，調用AdaDeltaPreSolve函數
  explicit AdaDeltaSolver(const SolverParameter& param) : SGDSolver<Dtype>(param) { AdaDeltaPreSolve(); }
  explicit AdaDeltaSolver(const string& param_file) : SGDSolver<Dtype>(param_file) { AdaDeltaPreSolve(); }

 protected:
  void AdaDeltaPreSolve();
// 計算並更新對應Blob值
  virtual void ComputeUpdateValue(int param_id, Dtype rate);

// 禁止使用AdaDeltaSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(AdaDeltaSolver);
};

/**
 * @brief AdamSolver, an algorithm for first-order gradient-based optimization
 *        of stochastic objective functions, based on adaptive estimates of
 *        lower-order moments. Described in [1].
 *
 * [1] D. P. Kingma and J. L. Ba, "ADAM: A Method for Stochastic Optimization."
 *     arXiv preprint arXiv:1412.6980v8 (2014).
 */
template <typename Dtype>
class AdamSolver : public SGDSolver<Dtype> { // 模板類AdamSolver。繼承SGDSolver
 public:
// 顯示構造函數，調用AdamPreSolve函數
  explicit AdamSolver(const SolverParameter& param) : SGDSolver<Dtype>(param) { AdamPreSolve();}
  explicit AdamSolver(const string& param_file) : SGDSolver<Dtype>(param_file) { AdamPreSolve(); }

 protected:
  void AdamPreSolve();
// 計算並更新對應Blob值
  virtual void ComputeUpdateValue(int param_id, Dtype rate);

// 禁止使用AdamSolver類的拷貝和賦值操作
  DISABLE_COPY_AND_ASSIGN(AdamSolver);
};

// new一個指定的solver方法對象
template <typename Dtype>
Solver<Dtype>* GetSolver(const SolverParameter& param) {
  SolverParameter_SolverType type = param.solver_type();

  switch (type) {
  case SolverParameter_SolverType_SGD:
      return new SGDSolver<Dtype>(param);
  case SolverParameter_SolverType_NESTEROV:
      return new NesterovSolver<Dtype>(param);
  case SolverParameter_SolverType_ADAGRAD:
      return new AdaGradSolver<Dtype>(param);
  case SolverParameter_SolverType_RMSPROP:
      return new RMSPropSolver<Dtype>(param);
  case SolverParameter_SolverType_ADADELTA:
      return new AdaDeltaSolver<Dtype>(param);
  case SolverParameter_SolverType_ADAM:
      return new AdamSolver<Dtype>(param);
  default:
      LOG(FATAL) << "Unknown SolverType: " << type;
  }
  return (Solver<Dtype>*) NULL;
}

}  // namespace caffe

#endif  // CAFFE_OPTIMIZATION_SOLVER_HPP_

// NOTE
// Update the next available ID when you add a new SolverParameter field.
//
// SolverParameter next available ID: 40 (last added: momentum2)
message SolverParameter { // Solver參數
  //////////////////////////////////////////////////////////////////////////////
  // Specifying the train and test networks
  //
  // Exactly one train net must be specified using one of the following fields:
  //     train_net_param, train_net, net_param, net
  // One or more test nets may be specified using any of the following fields:
  //     test_net_param, test_net, net_param, net
  // If more than one test net field is specified (e.g., both net and
  // test_net are specified), they will be evaluated in the field order given
  // above: (1) test_net_param, (2) test_net, (3) net_param/net.
  // A test_iter must be specified for each test_net.
  // A test_level and/or a test_stage may also be specified for each test_net.
  //////////////////////////////////////////////////////////////////////////////

  // Proto filename for the train net, possibly combined with one or more test nets.
  optional string net = 24; // .prototxt文件名稱， train or test net
  // Inline train net param, possibly combined with one or more test nets.
  optional NetParameter net_param = 25; // net parameter類

  optional string train_net = 1; // Proto filename for the train net, .prototxt文件名稱，train net
  repeated string test_net = 2; // Proto filenames for the test nets, .prototxt文件名稱，test net
  optional NetParameter train_net_param = 21; // Inline train net params, train net parameter類
  repeated NetParameter test_net_param = 22; // Inline test net params, test net parameter類

  // The states for the train/test nets. Must be unspecified or
  // specified once per net.
  //
  // By default, all states will have solver = true;
  // train_state will have phase = TRAIN,
  // and all test_state‘s will have phase = TEST.
  // Other defaults are set according to the NetState defaults.
  optional NetState train_state = 26; // train net state
  repeated NetState test_state = 27; // test net state

  // The number of iterations for each test net.
  repeated int32 test_iter = 3; // 對於測試網絡(用於評估)運行一次須要叠代的次數, test_iter * batch_size = 測試圖像總數量

  // The number of iterations between two testing phases.
  optional int32 test_interval = 4 [default = 0];　// 指定運行多少次訓練網絡運行一次測試網絡
  optional bool test_compute_loss = 19 [default = false]; // 運行測試網絡時是否計算loss
  // If true, run an initial test pass before the first iteration,
  // ensuring memory availability and printing the starting value of the loss.
  optional bool test_initialization = 32 [default = true]; // 在總的開始前，是否先運行一次測試網絡
  optional float base_lr = 5; // The base learning rate,基礎學習率
  // the number of iterations between displaying info. If display = 0, no info
  // will be displayed.
  optional int32 display = 6; // 指定叠代多少次顯示一次結果信息
  // Display the loss averaged over the last average_loss iterations
  optional int32 average_loss = 33 [default = 1]; // 
  optional int32 max_iter = 7; // the maximum number of iterations
  // accumulate gradients over `iter_size` x `batch_size` instances
  optional int32 iter_size = 36 [default = 1]; // 

  // The learning rate decay policy. The currently implemented learning rate
  // policies are as follows: // 學習率衰減策略
  //    - fixed: always return base_lr.
  //    - step: return base_lr * gamma ^ (floor(iter / step))
  //    - exp: return base_lr * gamma ^ iter
  //    - inv: return base_lr * (1 + gamma * iter) ^ (- power)
  //    - multistep: similar to step but it allows non uniform steps defined by
  //      stepvalue
  //    - poly: the effective learning rate follows a polynomial decay, to be
  //      zero by the max_iter. return base_lr (1 - iter/max_iter) ^ (power)
  //    - sigmoid: the effective learning rate follows a sigmod decay
  //      return base_lr ( 1/(1 + exp(-gamma * (iter - stepsize))))
  //
  // where base_lr, max_iter, gamma, step, stepvalue and power are defined
  // in the solver parameter protocol buffer, and iter is the current iteration.
  optional string lr_policy = 8; // 學習策略,可取的值包含：fixed、step、exp、inv、multistep、poly、sigmoid
  optional float gamma = 9; // The parameter to compute the learning rate.
  optional float power = 10; // The parameter to compute the learning rate.
  optional float momentum = 11; // The momentum value, 動量
  optional float weight_decay = 12; // The weight decay. // 
  // regularization types supported: L1 and L2
  // controlled by weight_decay
  optional string regularization_type = 29 [default = "L2"]; // L1 or L2
  // the stepsize for learning rate policy "step"
  optional int32 stepsize = 13; //
  // the stepsize for learning rate policy "multistep"
  repeated int32 stepvalue = 34; //

  // Set clip_gradients to >= 0 to clip parameter gradients to that L2 norm,
  // whenever their actual L2 norm is larger.
  optional float clip_gradients = 35 [default = -1]; //

  optional int32 snapshot = 14 [default = 0]; // The snapshot interval, 叠代多少次保存下結果(如權值、偏置)
  optional string snapshot_prefix = 15; // The prefix for the snapshot,指定保存文件名稱的前綴
  // whether to snapshot diff in the results or not. Snapshotting diff will help
  // debugging but the final protocol buffer size will be much larger.
  optional bool snapshot_diff = 16 [default = false]; //
  enum SnapshotFormat {
    HDF5 = 0;
    BINARYPROTO = 1;
  }
  optional SnapshotFormat snapshot_format = 37 [default = BINARYPROTO]; // HDF5 or BINARYPROTO
  // the mode solver will use: 0 for CPU and 1 for GPU. Use GPU in default.
  enum SolverMode {
    CPU = 0;
    GPU = 1;
  }
  optional SolverMode solver_mode = 17 [default = GPU]; // 指定solve mode是CPU還是GPU
  // the device_id will that be used in GPU mode. Use device_id = 0 in default.
  optional int32 device_id = 18 [default = 0]; // GPU mode下使用
  // If non-negative, the seed with which the Solver will initialize the Caffe
  // random number generator -- useful for reproducible results. Otherwise,
  // (and by default) initialize using a seed derived from the system clock.
  optional int64 random_seed = 20 [default = -1]; // 

  // Solver type
  enum SolverType { // solver優化方法
    SGD = 0;
    NESTEROV = 1;
    ADAGRAD = 2;
    RMSPROP = 3;
    ADADELTA = 4;
    ADAM = 5;
  }
  optional SolverType solver_type = 30 [default = SGD]; // 指定solver優化方法
  // numerical stability for RMSProp, AdaGrad and AdaDelta and Adam
  optional float delta = 31 [default = 1e-8]; // 
  // parameters for the Adam solver
  optional float momentum2 = 39 [default = 0.999]; // 

  // RMSProp decay value
  // MeanSquare(t) = rms_decay*MeanSquare(t-1) + (1-rms_decay)*SquareGradient(t)
  optional float rms_decay = 38; // 

  // If true, print information about the state of the net that may help with
  // debugging learning problems.
  optional bool debug_info = 23 [default = false]; // 

  // If false, don‘t save a snapshot after training finishes.
  optional bool snapshot_after_train = 28 [default = true]; //
}

#include "funset.hpp"
#include <string>
#include <vector>
#include <map>
#include "common.hpp"

int test_caffe_solver()
{
	caffe::Caffe::set_mode(caffe::Caffe::CPU); // set run caffe mode

	const std::string solver_prototxt{ "E:/GitCode/Caffe_Test/test_data/model/mnist/lenet_solver.prototxt" };

	caffe::SolverParameter solver_param;
	if (!caffe::ReadProtoFromTextFile(solver_prototxt.c_str(), &solver_param)) {
		fprintf(stderr, "parse solver.prototxt fail\n");
		return -1;
	}

	boost::shared_ptr<caffe::Solver<float> > solver(caffe::GetSolver<float>(solver_param));

	caffe::SolverParameter param = solver->param();

	if (param.has_net())
		fprintf(stderr, "net: %s\n", param.net().c_str());
	if (param.has_net_param()) {
		fprintf(stderr, "has net param\n");
		caffe::NetParameter net_param = param.net_param();
		if (net_param.has_name())
			fprintf(stderr, "net param name: %s\n", net_param.name().c_str());
	}
	if (param.has_train_net())
		fprintf(stderr, "train_net: %s\n", param.train_net());
	if (param.test_net().size() > 0) {
		for (auto test_net : param.test_net())
			fprintf(stderr, "test_net: %s\n", test_net);
	}
	if (param.has_train_net_param()) {
		fprintf(stderr, "has train net param\n");
		caffe::NetParameter train_net_param = param.train_net_param();
	}
	if (param.test_net_param().size() > 0) {
		fprintf(stderr, "has test net param\n");
		std::vector<caffe::NetParameter> test_net_param;
		for (auto net_param : param.test_net_param())
			test_net_param.push_back(net_param);
	}

	if (param.has_train_state()) {
		fprintf(stderr, "has train state\n");
		caffe::NetState state = param.train_state();
	}
	if (param.test_state().size()) {
		fprintf(stderr, "has test state\n");
	}

	if (param.test_iter_size() > 0) {
		fprintf(stderr, "has test iter\n");
		for (auto iter : param.test_iter())
			fprintf(stderr, "  %d  ", iter);
		fprintf(stderr, "\n");
	}

	if (param.has_test_interval())
		fprintf(stderr, "test interval: %d\n", param.test_interval());
	bool test_compute_loss = param.test_compute_loss();
	fprintf(stderr, "test compute loss: %d\n", test_compute_loss);
	bool test_initialization = param.test_initialization();
	fprintf(stderr, "test initializtion: %d\n", test_initialization);
	if (param.has_base_lr()) {
		float base_lr = param.base_lr();
		fprintf(stderr, "base lr: %f\n", base_lr);
	}
	if (param.has_display()) {
		int display = param.display();
		fprintf(stderr, "display: %d\n", display);
	}
	int average_loss = param.average_loss();
	fprintf(stderr, "average loss: %d\n", average_loss);
	if (param.has_max_iter()) {
		int max_iter = param.max_iter();
		fprintf(stderr, "max iter: %d\n", max_iter);
	}
	int iter_size = param.iter_size();
	fprintf(stderr, "iter size: %d\n", iter_size);

	if (param.has_lr_policy())
		fprintf(stderr, "lr policy: %s\n", param.lr_policy().c_str());
	if (param.has_gamma())
		fprintf(stderr, "gamma: %f\n", param.gamma());
	if (param.has_power())
		fprintf(stderr, "power: %f\n", param.power());
	if (param.has_momentum())
		fprintf(stderr, "momentum: %f\n", param.momentum());
	if (param.has_weight_decay())
		fprintf(stderr, "weight decay: %f\n", param.weight_decay());
	std::string regularization_type = param.regularization_type();
	fprintf(stderr, "regularization type: %s\n", param.regularization_type().c_str());
	if (param.has_stepsize())
		fprintf(stderr, "stepsize: %d\n", param.stepsize());
	if (param.stepvalue_size() > 0) {
		fprintf(stderr, "has stepvalue\n");
		for (auto value : param.stepvalue())
			fprintf(stderr, "  %d  ", value);
		fprintf(stderr, "\n");
	}

	fprintf(stderr, "clip gradients: %f\n", param.clip_gradients());

	fprintf(stderr, "snapshot: %d\n", param.snapshot());
	if (param.has_snapshot_prefix())
		fprintf(stderr, "snapshot prefix: %s\n", param.snapshot_prefix().c_str());
	fprintf(stderr, "snapshot diff: %d\n", param.snapshot_diff());
	caffe::SolverParameter_SnapshotFormat snapshot_format = param.snapshot_format();
	fprintf(stderr, "snapshot format: %s\n", snapshot_format == 0 ?
 "HDF5" : "BINARYPROTO");
	caffe::SolverParameter_SolverMode solver_mode = param.solver_mode();
	fprintf(stderr, "solver mode: %s\n", solver_mode == 0 ? "CPU" : "GPU");
	if (param.has_device_id())
		fprintf(stderr, "device id: %d\n", param.device_id());
	fprintf(stderr, "random seed: %d\n", param.random_seed());

	caffe::SolverParameter_SolverType solver_type = param.solver_type();
	std::string solver_method[] {"SGD", "NESTEROV", "ADAGRAD", "RMSPROP", "ADADELTA", "ADAM"};
	fprintf(stderr, "solver type: %s\n", solver_method[solver_type].c_str());
	fprintf(stderr, "delta: %f\n", param.delta());
	fprintf(stderr, "momentum2: %f\n", param.momentum2());

	if (param.has_rms_decay())
		fprintf(stderr, "rms decy: %f\n", param.rms_decay());

	fprintf(stderr, "debug info: %d\n", param.debug_info());
	fprintf(stderr, "snapshot after train: %d\n", param.snapshot_after_train());

	boost::shared_ptr<caffe::Net<float>> net = solver->net();
	std::vector<boost::shared_ptr<caffe::Net<float>>> test_nets = solver->test_nets();
	fprintf(stderr, "test nets size: %d\n", test_nets.size());
	fprintf(stderr, "iter: %d\n", solver->iter());

	return 0;
}

GitHub：https://github.com/fengbingchun/Caffe_Test

Caffe源代碼中Solver文件分析