C ++ tensorflow：打印节点名称时出错答案

【问题标题】：C++ tensorflow: error while printing node namesC ++ tensorflow：打印节点名称时出错
【发布时间】：2017-07-19 16:39:28
【问题描述】：

我正在尝试使用 tensorflow c++ 导入 frozen_model.pb。我想检查图中的所有节点是否正确导入。我找到了示例here，我尝试使用它来打印节点名称。我正在使用以下代码 sn-p。

int node_count = graph_def.node_size();
for (int i = 0; i < node_count; i++) {
    auto n = graph_def.node(i); // <-- error here
    cout<< n.name();
}

一旦带有注释的行（此处为错误）运行。它给我一个奇怪的错误。这是完整的错误。

CMakeFiles/c++.dir/main.cc.o: In function `google::protobuf::RepeatedPtrField<tensorflow::NodeDef>::TypeHandler::Type const& google::protobuf::internal::RepeatedPtrFieldBase::Get<google::protobuf::RepeatedPtrField<tensorflow::NodeDef>::TypeHandler>(int) const':
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x3a): undefined reference to `google::protobuf::internal::LogMessage::LogMessage(google::protobuf::LogLevel, char const*, int)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x50): undefined reference to `google::protobuf::internal::LogMessage::operator<<(char const*)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x68): undefined reference to `google::protobuf::internal::LogFinisher::operator=(google::protobuf::internal::LogMessage&)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x7c): undefined reference to `google::protobuf::internal::LogMessage::~LogMessage()'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0xb1): undefined reference to `google::protobuf::internal::LogMessage::LogMessage(google::protobuf::LogLevel, char const*, int)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0xc7): undefined reference to `google::protobuf::internal::LogMessage::operator<<(char const*)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0xdf): undefined reference to `google::protobuf::internal::LogFinisher::operator=(google::protobuf::internal::LogMessage&)'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0xf3): undefined reference to `google::protobuf::internal::LogMessage::~LogMessage()'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x12a): undefined reference to `google::protobuf::internal::LogMessage::~LogMessage()'
main.cc:(.text._ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi[_ZNK6google8protobuf8internal20RepeatedPtrFieldBase3GetINS0_16RepeatedPtrFieldIN10tensorflow7NodeDefEE11TypeHandlerEEERKNT_4TypeEi]+0x14f): undefined reference to `google::protobuf::internal::LogMessage::~LogMessage()'
collect2: error: ld returned 1 exit status
make[2]: *** [../bin/c++] Error 1
make[1]: *** [CMakeFiles/c++.dir/all] Error 2
make: *** [all] Error 2

我得到了正确的node_size 编号，即21。你可以在这里看到我在python中导入并打印了同一个frozen_graph.pb的节点

【问题讨论】：

你是如何构建这个程序的？看起来libprotobuf 尚未链接。
我使用此存储库github.com/cjweeks/tensorflow-cmake 设置项目，因为我想使用 cmake 构建 c++ 项目

标签： c++ tensorflow

【解决方案1】：

在@mrry 评论的帮助下。我最终通过在repository 中包含FindProtoBuf.cmake 文件，在我的CMakeLists.txt 中手动链接libprotobuf。我复制到与CMakeLists.txt 相同的目录中，并按照this 中的命令回答如何在cmake 中包含ProtoBuf。

【讨论】：