Caffe源码分析—phase/level/stage理解

在caffe中，train.prototxt 、test.prototxt和 deploy.prototxt中间为了增加网络结构（.prototxt）文件的灵活性，可以通过设置phase、stage、level参数实现train、test、deploy文件统一合并在一个.prototxt文件中（通常见到的train_test合并为一个文件或分开为两个文件就是使用了这些参数）。

1. 对于一个prototxt文件，可以在NetParameter层面，通过state字段来表示当前proto文件的phase/level/stage信息。

message NetParameter {
...
  // The current "state" of the network, including the phase, level, and stage.
  // Some layers may be included/excluded depending on this state and the states
  // specified in the layers' include and exclude fields.
  optional NetState state = 6;
...
  // The layers that make up the net.  
  repeated LayerParameter layer = 100;  
}
message NetState {
  optional Phase phase = 1 [default = TEST];
  optional int32 level = 2 [default = 0];
  repeated string stage = 3;
}

其中phase指定当前Proto文件运行所在阶段，level要求Layer级别以及stage过滤。通过这三个字段可以选择或过滤掉相关的Layer。

2. 对于每一个layer，定义了两个NetStateRule类型的字段，分别是include和exclude，来控制该Layer在Netstat处于何种状态下存在。

message LayerParameter {
  // Rules controlling whether and when a layer is included in the network,
  // based on the current NetState.  You may specify a non-zero number of rules
  // to include OR exclude, but not both.  If no include or exclude rules are
  // specified, the layer is always included.  If the current NetState meets
  // ANY (i.e., one or more) of the specified rules, the layer is
  // included/excluded.
  repeated NetStateRule include = 8;
  repeated NetStateRule exclude = 9;
}
message NetStateRule {
  // Set phase to require the NetState have a particular phase (TRAIN or TEST)
  // to meet this rule.
  optional Phase phase = 1;

  // Set the minimum and/or maximum levels in which the layer should be used.
  // Leave undefined to meet the rule regardless of level.
  optional int32 min_level = 2;
  optional int32 max_level = 3;

  // Customizable sets of stages to include or exclude.
  // The net must have ALL of the specified stages and NONE of the specified
  // "not_stage"s to meet the rule.
  // (Use multiple NetStateRules to specify conjunctions of stages.)
  repeated string stage = 4;
  repeated string not_stage = 5;
}

3. NetParameter除了作为Net内部配置以外，与直接作为SolverParameter的字段来影响slover过程。SolverParameter定义还是比较复杂，但是整体目标是从SolverParameter中创建出唯一训练网络以及多个测试网络。

• SolverParameter中通过net, net_param, train_net, train_net_param四个字段来指定训练网络，但是只允许同时出现一个。其中net，train_net为Net文件，而net_param, train_net_param为内联的网络配置，其实本质上没有太大区别，读取到内存以后，都是NetParameter。
• SolverParameter中通过net, net_param, test_net, test_net_param四个字段来指定测试网络，其中net, net_param由于也会被读取为训练网络，所以只允许有一个存在。test_net, test_net_param运行同时存在并可以多于1个。但是最终有多少个测试网络实际运行是通过test_iter参数个数来确定的，如果只指定了2个test_iter，哪怕我们定义了3个测试网络，实际也只会运行2个，另外一个迭代次数为0。优先级上test_net_param>test_net>net_param>net。
• 对于训练网络，可以在SolverParameter中配置train_state，train_state内容会和网络NetParameter内部的Netstate进行合并，其中train_state数据的优先级要高于NetParameter内部的Netstate。
• 对于测试网络，要么给每一个测试网络都在SolverParameter中配置一个test_state，要么都不配置。每一个测试网络的test_state会和测试网络NetParameter内部的Netstate进行合并，其中test_state数据的优先级高于NetParameter内部的Netstate。
• 对于训练网络，Netstat.phase恒等于train；对于测试网络，Netstate.phase恒等于test。

4. Caffe.cpp启动参数也支持设置state

DEFINE_string(phase, "",
    "Optional; network phase (TRAIN or TEST). Only used for 'time'.");
DEFINE_int32(level, 0,
    "Optional; network level.");
DEFINE_string(stage, "",
    "Optional; network stages (not to be confused with phase), "
    "separated by ','.");

这两个参数实际是与上面SolverParameter的train_state进行合并

  solver_param.mutable_train_state()->set_level(FLAGS_level);
 for (int i = 0; i < stages.size(); i++) {
    solver_param.mutable_train_state()->add_stage(stages[i]);
  }