зеркало из https://github.com/microsoft/caffe.git
Organize the loss layers in alphabetical order
This commit is contained in:
Kai Li
Родитель
8e1689faa5
Коммит
97deb9101b
|
|
@ -22,6 +22,36 @@ namespace caffe {
|
|||
|
||||
const float kLOG_THRESHOLD = 1e-20;
|
||||
|
||||
/* AccuracyLayer
|
||||
Note: not an actual loss layer! Does not implement backwards step.
|
||||
Computes the accuracy of argmax(a) with respect to b.
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class AccuracyLayer : public Layer<Dtype> {
|
||||
public:
|
||||
explicit AccuracyLayer(const LayerParameter& param)
|
||||
: Layer<Dtype>(param) {}
|
||||
virtual void SetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_ACCURACY;
|
||||
}
|
||||
|
||||
virtual inline int ExactNumBottomBlobs() const { return 2; }
|
||||
virtual inline int ExactNumTopBlobs() const { return 1; }
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom) {
|
||||
NOT_IMPLEMENTED;
|
||||
}
|
||||
|
||||
int top_k_;
|
||||
};
|
||||
|
||||
/* LossLayer
|
||||
Takes two inputs of same num (a and b), and has no output.
|
||||
The gradient is propagated to a.
|
||||
|
|
@ -45,6 +75,139 @@ class LossLayer : public Layer<Dtype> {
|
|||
}
|
||||
};
|
||||
|
||||
/* EuclideanLossLayer
|
||||
Compute the L_2 distance between the two inputs.
|
||||
|
||||
loss = (1/2 \sum_i (a_i - b_i)^2)
|
||||
a' = 1/I (a - b)
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class EuclideanLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit EuclideanLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param), diff_() {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_EUCLIDEAN_LOSS;
|
||||
}
|
||||
// Unlike most loss layers, in the EuclideanLossLayer we can backpropagate
|
||||
// to both inputs.
|
||||
virtual inline bool AllowForceBackward(const int bottom_index) const {
|
||||
return true;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual Dtype Forward_gpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
Blob<Dtype> diff_;
|
||||
};
|
||||
|
||||
/* HingeLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class HingeLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit HingeLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param) {}
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_HINGE_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
};
|
||||
|
||||
/* InfogainLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class InfogainLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit InfogainLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param), infogain_() {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_INFOGAIN_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
Blob<Dtype> infogain_;
|
||||
};
|
||||
|
||||
/* MultinomialLogisticLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class MultinomialLogisticLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit MultinomialLogisticLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param) {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_MULTINOMIAL_LOGISTIC_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
};
|
||||
|
||||
/* SigmoidCrossEntropyLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class SigmoidCrossEntropyLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit SigmoidCrossEntropyLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param),
|
||||
sigmoid_layer_(new SigmoidLayer<Dtype>(param)),
|
||||
sigmoid_output_(new Blob<Dtype>()) {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_SIGMOID_CROSS_ENTROPY_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual Dtype Forward_gpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
shared_ptr<SigmoidLayer<Dtype> > sigmoid_layer_;
|
||||
// sigmoid_output stores the output of the sigmoid layer.
|
||||
shared_ptr<Blob<Dtype> > sigmoid_output_;
|
||||
// Vector holders to call the underlying sigmoid layer forward and backward.
|
||||
vector<Blob<Dtype>*> sigmoid_bottom_vec_;
|
||||
vector<Blob<Dtype>*> sigmoid_top_vec_;
|
||||
};
|
||||
|
||||
// Forward declare SoftmaxLayer for use in SoftmaxWithLossLayer.
|
||||
template <typename Dtype> class SoftmaxLayer;
|
||||
|
||||
|
|
@ -92,169 +255,6 @@ class SoftmaxWithLossLayer : public Layer<Dtype> {
|
|||
vector<Blob<Dtype>*> softmax_top_vec_;
|
||||
};
|
||||
|
||||
/* SigmoidCrossEntropyLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class SigmoidCrossEntropyLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit SigmoidCrossEntropyLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param),
|
||||
sigmoid_layer_(new SigmoidLayer<Dtype>(param)),
|
||||
sigmoid_output_(new Blob<Dtype>()) {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_SIGMOID_CROSS_ENTROPY_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual Dtype Forward_gpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
shared_ptr<SigmoidLayer<Dtype> > sigmoid_layer_;
|
||||
// sigmoid_output stores the output of the sigmoid layer.
|
||||
shared_ptr<Blob<Dtype> > sigmoid_output_;
|
||||
// Vector holders to call the underlying sigmoid layer forward and backward.
|
||||
vector<Blob<Dtype>*> sigmoid_bottom_vec_;
|
||||
vector<Blob<Dtype>*> sigmoid_top_vec_;
|
||||
};
|
||||
|
||||
/* EuclideanLossLayer
|
||||
Compute the L_2 distance between the two inputs.
|
||||
|
||||
loss = (1/2 \sum_i (a_i - b_i)^2)
|
||||
a' = 1/I (a - b)
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class EuclideanLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit EuclideanLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param), diff_() {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_EUCLIDEAN_LOSS;
|
||||
}
|
||||
// Unlike most loss layers, in the EuclideanLossLayer we can backpropagate
|
||||
// to both inputs.
|
||||
virtual inline bool AllowForceBackward(const int bottom_index) const {
|
||||
return true;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual Dtype Forward_gpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
virtual void Backward_gpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
Blob<Dtype> diff_;
|
||||
};
|
||||
|
||||
/* InfogainLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class InfogainLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit InfogainLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param), infogain_() {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_INFOGAIN_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
|
||||
Blob<Dtype> infogain_;
|
||||
};
|
||||
|
||||
/* HingeLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class HingeLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit HingeLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param) {}
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_HINGE_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
};
|
||||
|
||||
/* MultinomialLogisticLossLayer
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class MultinomialLogisticLossLayer : public LossLayer<Dtype> {
|
||||
public:
|
||||
explicit MultinomialLogisticLossLayer(const LayerParameter& param)
|
||||
: LossLayer<Dtype>(param) {}
|
||||
virtual void FurtherSetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_MULTINOMIAL_LOGISTIC_LOSS;
|
||||
}
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom);
|
||||
};
|
||||
|
||||
/* AccuracyLayer
|
||||
Note: not an actual loss layer! Does not implement backwards step.
|
||||
Computes the accuracy of argmax(a) with respect to b.
|
||||
*/
|
||||
template <typename Dtype>
|
||||
class AccuracyLayer : public Layer<Dtype> {
|
||||
public:
|
||||
explicit AccuracyLayer(const LayerParameter& param)
|
||||
: Layer<Dtype>(param) {}
|
||||
virtual void SetUp(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
|
||||
virtual inline LayerParameter_LayerType type() const {
|
||||
return LayerParameter_LayerType_ACCURACY;
|
||||
}
|
||||
|
||||
virtual inline int ExactNumBottomBlobs() const { return 2; }
|
||||
virtual inline int ExactNumTopBlobs() const { return 1; }
|
||||
|
||||
protected:
|
||||
virtual Dtype Forward_cpu(const vector<Blob<Dtype>*>& bottom,
|
||||
vector<Blob<Dtype>*>* top);
|
||||
virtual void Backward_cpu(const vector<Blob<Dtype>*>& top,
|
||||
const vector<bool>& propagate_down, vector<Blob<Dtype>*>* bottom) {
|
||||
NOT_IMPLEMENTED;
|
||||
}
|
||||
|
||||
int top_k_;
|
||||
};
|
||||
|
||||
} // namespace caffe
|
||||
|
||||
#endif // CAFFE_LOSS_LAYERS_HPP_
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче