ceras::compiled_model< Model, Optimizer, Loss > Struct Template Reference

#include <model.hpp>

Public Types
typedef Model::input_layer_type	io_layer_type

Public Member Functions
	compiled_model (Model const &m, io_layer_type const &input_place_holder, io_layer_type const &ground_truth_place_holder, Loss const &loss, Optimizer const &optimizer)
template<Tensor Tsor>
auto	evaluate (Tsor const &inputs, Tsor const &outputs, unsigned long batch_size=32)
template<Tensor Tsor>
auto	fit (Tsor const &inputs, Tsor const &outputs, unsigned long batch_size, unsigned long epoch=1, int verbose=0, double validation_split=0.0)
template<Tensor Tsor>
auto	train_on_batch (Tsor const &input, Tsor const &output)
template<Tensor Tsor>
auto	predict (Tsor const &input_tensor)
template<Expression Exp>
auto	operator() (Exp const &ex) const noexcept
void	trainable (bool t)

Public Attributes
decltype(std::declval< Optimizer >()(std::declval< Loss & >())) typedef	optimizer_type
Model	model_
io_layer_type	input_place_holder_
io_layer_type	ground_truth_place_holder_
Loss	loss_
Optimizer	optimizer_
optimizer_type	compiled_optimizer_

Member Typedef Documentation

io_layer_type

template<typename Model , typename Optimizer , typename Loss >

typedef Model::input_layer_type ceras::compiled_model< Model, Optimizer, Loss >::io_layer_type

Constructor & Destructor Documentation

compiled_model()

template<typename Model , typename Optimizer , typename Loss >

ceras::compiled_model< Model, Optimizer, Loss >::compiled_model ( Model const &  m, io_layer_type const &  input_place_holder, io_layer_type const &  ground_truth_place_holder, Loss const &  loss, Optimizer const &  optimizer  )

inline

Member Function Documentation

evaluate()

template<typename Model , typename Optimizer , typename Loss >

template<Tensor Tsor>

auto ceras::compiled_model< Model, Optimizer, Loss >::evaluate ( Tsor const &  inputs, Tsor const &  outputs, unsigned long  batch_size = 32  )

inline

Calculate the loss for the model in test model.

Parameters

inputs	Input data. A tensor of shape (samples, input_shape).
outputs	Output data. A tensor of shape (samples, output_shape).
batch_size	Number of samples per batch of computation. Default to 32.

Returns

Test loss. A scalar.

fit()

template<typename Model , typename Optimizer , typename Loss >

template<Tensor Tsor>

auto ceras::compiled_model< Model, Optimizer, Loss >::fit ( Tsor const &  inputs, Tsor const &  outputs, unsigned long  batch_size, unsigned long  epoch = 1, int  verbose = 0, double  validation_split = 0.0  )

inline

Train the model on the selected dataset for a fixed numbers of epoches.

Parameters

inputs	Input data. A tensor of shape (samples, input_shape).
outputs	Input data. A tensor of shape (samples, output_shape).
batch_size	Number of samples per gradient update. Should agree with the batch size in the optimizer.
epoch	Number of epoches to train the dataset.
verbose	Verbosity mode. 0 for slient. 1 for one line per epoch.
validation_split	Fraction of the training data that will be used for validation. A floating number in range [0, 1].

Returns

A tuple of two vectors. The first vector gives the historical errors on the training data. The second vector gives the historical errors on the validation data.

Example:

model m{ ... };
auto cm = m.compile( ... );
tensor<float> inputs, outputs;
//...
unsigned long batch_size = 32;
unsigned long epoch = 10;
int verbose = 1;
double validation_split = 0.2;
auto errors = cm.fit( inputs, outputs, batch_size, epoch, verbose, validation_split );

operator()()

template<typename Model , typename Optimizer , typename Loss >

template<Expression Exp>

auto ceras::compiled_model< Model, Optimizer, Loss >::operator() ( Exp const &  ex ) const

inlinenoexcept

predict()

template<typename Model , typename Optimizer , typename Loss >

template<Tensor Tsor>

auto ceras::compiled_model< Model, Optimizer, Loss >::predict ( Tsor const &  input_tensor )

inline

train_on_batch()

template<typename Model , typename Optimizer , typename Loss >

template<Tensor Tsor>

auto ceras::compiled_model< Model, Optimizer, Loss >::train_on_batch ( Tsor const &  input, Tsor const &  output  )

inline

Running a single updated on a single batch of data.

Parameters

input	The input data to train the model. A tensor of shape (batch_size, input_shape).
output	The output data to train the model. A tensor of shape (batch_size, output_shape).

Returns

Training loss. A scalar.

Example code:

auto m = model{ ... };
auto cm = m.compile( ... );
for ( auto idx : range( 1024 ) )
{
    auto x = ...; // get batch input
    auto y = ...; // get batch output
    cm.train_on_batch( x, y );
}

trainable()

template<typename Model , typename Optimizer , typename Loss >

void ceras::compiled_model< Model, Optimizer, Loss >::trainable ( bool  t )

inline

Member Data Documentation

compiled_optimizer_

template<typename Model , typename Optimizer , typename Loss >

optimizer_type ceras::compiled_model< Model, Optimizer, Loss >::compiled_optimizer_

ground_truth_place_holder_

template<typename Model , typename Optimizer , typename Loss >

io_layer_type ceras::compiled_model< Model, Optimizer, Loss >::ground_truth_place_holder_

input_place_holder_

template<typename Model , typename Optimizer , typename Loss >

io_layer_type ceras::compiled_model< Model, Optimizer, Loss >::input_place_holder_

loss_

template<typename Model , typename Optimizer , typename Loss >

Loss ceras::compiled_model< Model, Optimizer, Loss >::loss_

model_

template<typename Model , typename Optimizer , typename Loss >

Model ceras::compiled_model< Model, Optimizer, Loss >::model_

optimizer_

template<typename Model , typename Optimizer , typename Loss >

Optimizer ceras::compiled_model< Model, Optimizer, Loss >::optimizer_

optimizer_type

template<typename Model , typename Optimizer , typename Loss >

decltype(std::declval<Optimizer>()(std::declval<Loss&>())) typedef ceras::compiled_model< Model, Optimizer, Loss >::optimizer_type

---

The documentation for this struct was generated from the following file:

• /home/feng/workspace/github.repo/ceras/include/model.hpp