Question

要清楚，我不想在C ++中加载预训练模型。我想用C ++进行培训，然后使用该模型。

Answer 1

我已经尝试使用C ++ API进行培训几天了，据我所知，使用1.2版本没有完整的示例。我还没有放弃，如果我设法做出一个有效的例子（我正在尝试识别来自MNIST的数字）我会在这里发布它。在1.2中，会话的运行方式发生了变化，文档非常缺乏。

如果您愿意使用旧版本的C ++ API，源代码中的这个示例就足够了：

https://github.com/tensorflow/tensorflow/blob/master/tensorflow/cc/tutorials/example_trainer.cc。

Answer 2

下面提供了一个很好的示例，可以从头开始训练模型，不幸的是，无法保存训练后的示例。他还解释了代码。我一直在用这个： https://matrices.io/training-a-deep-neural-network-using-only-tensorflow-c/ （这是在c ++中训练tensorflow-NN的第一个示例）

我的存储库中也进行了一些基本测试： https://github.com/PinkySan/TensorflowHandlingTests

  Scope scope = Scope::NewRootScope();

  auto x = Placeholder(scope, DT_FLOAT);
  auto y = Placeholder(scope, DT_FLOAT);

  // weights init
  auto w1 = Variable(scope, {3, 3}, DT_FLOAT);
  auto assign_w1 = Assign(scope, w1, RandomNormal(scope, {3, 3}, DT_FLOAT));

  auto w2 = Variable(scope, {3, 2}, DT_FLOAT);
  auto assign_w2 = Assign(scope, w2, RandomNormal(scope, {3, 2}, DT_FLOAT));

  auto w3 = Variable(scope, {2, 1}, DT_FLOAT);
  auto assign_w3 = Assign(scope, w3, RandomNormal(scope, {2, 1}, DT_FLOAT));

  // bias init
  auto b1 = Variable(scope, {1, 3}, DT_FLOAT);
  auto assign_b1 = Assign(scope, b1, RandomNormal(scope, {1, 3}, DT_FLOAT));

  auto b2 = Variable(scope, {1, 2}, DT_FLOAT);
  auto assign_b2 = Assign(scope, b2, RandomNormal(scope, {1, 2}, DT_FLOAT));

  auto b3 = Variable(scope, {1, 1}, DT_FLOAT);
  auto assign_b3 = Assign(scope, b3, RandomNormal(scope, {1, 1}, DT_FLOAT));

  // layers
  auto layer_1 = Tanh(scope, Tanh(scope, Add(scope, MatMul(scope, x, w1), b1)));
  auto layer_2 = Tanh(scope, Add(scope, MatMul(scope, layer_1, w2), b2));
  auto layer_3 = Tanh(scope, Add(scope, MatMul(scope, layer_2, w3), b3));

  // regularization
  auto regularization = AddN(scope,
                             initializer_list<Input>{L2Loss(scope, w1),
                                                     L2Loss(scope, w2),
                                                     L2Loss(scope, w3)});

  // loss calculation
  auto loss = Add(scope,
                  ReduceMean(scope, Square(scope, Sub(scope, layer_3, y)), {0, 1}),
                  Mul(scope, Cast(scope, 0.01,  DT_FLOAT), regularization));

  // add the gradients operations to the graph
  std::vector<Output> grad_outputs;
  TF_CHECK_OK(AddSymbolicGradients(scope, {loss}, {w1, w2, w3, b1, b2, b3}, &grad_outputs));

  // update the weights and bias using gradient descent
  auto apply_w1 = ApplyGradientDescent(scope, w1, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[0]});
  auto apply_w2 = ApplyGradientDescent(scope, w2, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[1]});
  auto apply_w3 = ApplyGradientDescent(scope, w3, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[2]});
  auto apply_b1 = ApplyGradientDescent(scope, b1, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[3]});
  auto apply_b2 = ApplyGradientDescent(scope, b2, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[4]});
  auto apply_b3 = ApplyGradientDescent(scope, b3, Cast(scope, 0.01,  DT_FLOAT), {grad_outputs[5]});

  ClientSession session(scope);
  std::vector<Tensor> outputs;

  // init the weights and biases by running the assigns nodes once
  TF_CHECK_OK(session.Run({assign_w1, assign_w2, assign_w3, assign_b1, assign_b2, assign_b3}, nullptr));

  // training steps
  for (int i = 0; i < 5000; ++i) {
    if (i % 100 == 0) {
      TF_CHECK_OK(session.Run({{x, x_data}, {y, y_data}}, {loss}, &outputs));
      std::cout << "Loss after " << i << " steps " << outputs[0].scalar<float>() << std::endl;
    }
    // nullptr because the output from the run is useless
    TF_CHECK_OK(session.Run({{x, x_data}, {y, y_data}}, {apply_w1, apply_w2, apply_w3, apply_b1, apply_b2, apply_b3}, nullptr));
  }

Answer 3

我认为最好的策略是在Python中训练模型，保存模型然后在C ++中加载冻结的模型。我写了说明，说明如何冻结Tensorflow（TFLearn）中的检查点并加载此模型以使用C / C ++ API进行推理：

https://github.com/kecsap/tensorflow_cpp_packaging/blob/master/README.md#save-a-checkpoint-in-python-tflearn

我将要复制的说明太长了。

是否有使用Tensorflow C ++ API的培训示例？

3 个答案: