Skip to content

D
dlib
Commit 2092e303 authored by Davis King's avatar Davis King
Browse files
Options

Renamed compute_loss() to compute_loss_value_and_gradient() in the loss 

interface.
parent ee2a0070
Hide whitespace changes
Inline Side-by-side
Showing with 23 additions and 21 deletions
dlib/dnn/core.h
View file @ 2092e303
......@@ -2171,7 +2171,7 @@ namespace dlib
{
subnetwork.forward(x);
dimpl::subnet_wrapper<subnet_type> wsub(subnetwork);
return loss.compute_loss(x, lbegin, wsub);
return loss.compute_loss_value_and_gradient(x, lbegin, wsub);
}
template <typename input_iterator, typename label_iterator>
......@@ -2191,7 +2191,7 @@ namespace dlib
{
subnetwork.forward(x);
dimpl::subnet_wrapper<subnet_type> wsub(subnetwork);
return loss.compute_loss(x, wsub);
return loss.compute_loss_value_and_gradient(x, wsub);
}
template <typename input_iterator>
......@@ -2212,7 +2212,7 @@ namespace dlib
{
subnetwork.forward(x);
dimpl::subnet_wrapper<subnet_type> wsub(subnetwork);
double l = loss.compute_loss(x, lbegin, wsub);
double l = loss.compute_loss_value_and_gradient(x, lbegin, wsub);
subnetwork.back_propagate_error(x);
return l;
}
......@@ -2232,7 +2232,7 @@ namespace dlib
{
subnetwork.forward(x);
dimpl::subnet_wrapper<subnet_type> wsub(subnetwork);
double l = loss.compute_loss(x, wsub);
double l = loss.compute_loss_value_and_gradient(x, wsub);
subnetwork.back_propagate_error(x);
return l;
}
......
dlib/dnn/loss.h
View file @ 2092e303
......@@ -47,7 +47,7 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
......@@ -148,7 +148,7 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
......@@ -259,7 +259,7 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
......
dlib/dnn/loss_abstract.h
View file @ 2092e303
......@@ -33,7 +33,8 @@ namespace dlib
Finally, note that there are two broad flavors of loss layer, supervised
and unsupervised. The EXAMPLE_LOSS_LAYER_ as shown here is a supervised
layer. To make an unsupervised loss you simply leave out the label_type
typedef, to_label(), and the truth iterator argument to compute_loss().
typedef, to_label(), and the truth iterator argument to
compute_loss_value_and_gradient().
!*/
public:
......@@ -90,7 +91,7 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
......@@ -116,9 +117,10 @@ namespace dlib
- This function computes a loss function that describes how well the output
of sub matches the expected labels given by truth. Let's write the loss
function as L(input_tensor, truth, sub).
- Then compute_loss() computes the gradient of L() with respect to the
outputs in sub. Specifically, compute_loss() assigns the gradients into
sub by performing the following tensor assignments, for all valid i:
- Then compute_loss_value_and_gradient() computes the gradient of L() with
respect to the outputs in sub. Specifically, compute_loss_value_and_gradient()
assigns the gradients into sub by performing the following tensor
assignments, for all valid i:
- layer<i>(sub).get_gradient_input() = the gradient of
L(input_tensor,truth,sub) with respect to layer<i>(sub).get_output().
- returns L(input_tensor,truth,sub)
......@@ -188,14 +190,14 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
) const;
/*!
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss() except
it has the additional calling requirements that:
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss_value_and_gradient()
except it has the additional calling requirements that:
- sub.get_output().nr() == 1
- sub.get_output().nc() == 1
- sub.get_output().k() == 1
......@@ -254,14 +256,14 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
) const;
/*!
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss() except
it has the additional calling requirements that:
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss_value_and_gradient()
except it has the additional calling requirements that:
- sub.get_output().nr() == 1
- sub.get_output().nc() == 1
- sub.get_output().k() == 1
......@@ -323,14 +325,14 @@ namespace dlib
typename const_label_iterator,
typename SUBNET
>
double compute_loss (
double compute_loss_value_and_gradient (
const tensor& input_tensor,
const_label_iterator truth,
SUBNET& sub
) const;
/*!
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss() except
it has the additional calling requirements that:
This function has the same interface as EXAMPLE_LOSS_LAYER_::compute_loss_value_and_gradient()
except it has the additional calling requirements that:
- sub.get_output().nr() == 1
- sub.get_output().nc() == 1
- sub.get_output().num_samples() == input_tensor.num_samples()
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment