MyCaffe  1.11.7.7
Deep learning software for Windows C# programmers.
MyCaffe.common.TrainingIterationArgs< T > Class Template Reference

The TrainingIterationArgs is sent to the Solver::OnTrainingIteration event that fires at the end of a training cycle. More...

Inheritance diagram for MyCaffe.common.TrainingIterationArgs< T >:
MyCaffe.common.TestingIterationArgs< T >

Public Member Functions

 TrainingIterationArgs (int nIteration, double dfAccuracy, double dfLoss, double dfSmoothedLoss, double dfBestSmoothedLoss, bool bWeightsUpdated, string strActiveLabelCounts, string strLabelQueryHitPercents, string strLabelQueryEpochs, string strBoostQueryHitPercents, double dfLearningRate, double dfMsTiming, DebugInformation< T > dbgInfo=null)
 The TrainingIterationArgs constructor. More...
 
- Public Member Functions inherited from MyCaffe.common.TestingIterationArgs< T >
 TestingIterationArgs (int nIteration, double dfAccuracy, double dfMsTiming)
 The TestingIterationArgs constructor. More...
 

Properties

double Loss [get]
 Returns the loss of the training cycle. More...
 
double SmoothedLoss [get]
 Retunrs the average loss after the training cycle. More...
 
double BestSmoothedLoss [get]
 Returns the best smoothed loss observed during the training. More...
 
bool WeightsUpdated [get]
 Returns whether or not the weights have been updated. More...
 
string ActiveLabelCounts [get]
 Returns the current active label counts as a string. More...
 
string LabelQueryHitPercents [get]
 Returns the current label query hit percentages observed at this point in training. More...
 
string LabelQueryEpochs [get]
 Returns the current label epochs per label observed at this point in training. More...
 
string BoostQueryHitPercents [get]
 Returns the current boost query hit percentages observed at this point in training. More...
 
double LearningRate [get]
 Return the current learning rate. More...
 
DebugInformation< T > DebugInformation [get]
 Returns the DebugInformation (if any). The DebugInformation is set to null when Solver::EnableBlobDebugging == false. More...
 
- Properties inherited from MyCaffe.common.TestingIterationArgs< T >
double Accuracy [get]
 Return the accuracy of the test cycle. More...
 
int Iteration [get]
 Return the iteration of the test cycle. More...
 
double Timing [get]
 Return the timing of the test cycle in ms. More...
 

Detailed Description

The TrainingIterationArgs is sent to the Solver::OnTrainingIteration event that fires at the end of a training cycle.

Template Parameters
TSpecifies the base type float or double. Using float is recommended to conserve GPU memory.

Definition at line 263 of file EventArgs.cs.

Constructor & Destructor Documentation

◆ TrainingIterationArgs()

MyCaffe.common.TrainingIterationArgs< T >.TrainingIterationArgs ( int  nIteration,
double  dfAccuracy,
double  dfLoss,
double  dfSmoothedLoss,
double  dfBestSmoothedLoss,
bool  bWeightsUpdated,
string  strActiveLabelCounts,
string  strLabelQueryHitPercents,
string  strLabelQueryEpochs,
string  strBoostQueryHitPercents,
double  dfLearningRate,
double  dfMsTiming,
DebugInformation< T >  dbgInfo = null 
)

The TrainingIterationArgs constructor.

Parameters
nIterationSpecifies the iteration of the training cycle.
dfAccuracySpecifies the last accuracy recieved during the training cycle's last testing cycle.
dfLossSpecifies the loss of the training cycle.
dfSmoothedLossSpecifies the averaged loss of the training cycle.
dfBestSmoothedLossSpecifies the best smoothed loss observed so far during the training.
bWeightsUpdatedSpecifies whether or not the weights have been updated.
strActiveLabelCountsSpecifies the current active label counts observed.
strLabelQueryHitPercentsSpecifies the current label query hit percentages observed.
strLabelQueryEpochsSpecifies the current label epoch count per label observed.
strBoostQueryHitPercentsSpecifies the current boost query hit percentages observed.
dfLearningRateSpecifies the current learning rate.
dfMsTimingSpecifies the timing of the training cycle.
dbgInfoOptionally, specifies the DebugInformation of the training cycle. This value is set when Solver::EnableBlobDebugging == true.

Definition at line 292 of file EventArgs.cs.

Property Documentation

◆ ActiveLabelCounts

string MyCaffe.common.TrainingIterationArgs< T >.ActiveLabelCounts
get

Returns the current active label counts as a string.

Definition at line 342 of file EventArgs.cs.

◆ BestSmoothedLoss

double MyCaffe.common.TrainingIterationArgs< T >.BestSmoothedLoss
get

Returns the best smoothed loss observed during the training.

Definition at line 326 of file EventArgs.cs.

◆ BoostQueryHitPercents

string MyCaffe.common.TrainingIterationArgs< T >.BoostQueryHitPercents
get

Returns the current boost query hit percentages observed at this point in training.

Definition at line 366 of file EventArgs.cs.

◆ DebugInformation

Returns the DebugInformation (if any). The DebugInformation is set to null when Solver::EnableBlobDebugging == false.

Definition at line 382 of file EventArgs.cs.

◆ LabelQueryEpochs

string MyCaffe.common.TrainingIterationArgs< T >.LabelQueryEpochs
get

Returns the current label epochs per label observed at this point in training.

Definition at line 358 of file EventArgs.cs.

◆ LabelQueryHitPercents

string MyCaffe.common.TrainingIterationArgs< T >.LabelQueryHitPercents
get

Returns the current label query hit percentages observed at this point in training.

Definition at line 350 of file EventArgs.cs.

◆ LearningRate

double MyCaffe.common.TrainingIterationArgs< T >.LearningRate
get

Return the current learning rate.

Definition at line 374 of file EventArgs.cs.

◆ Loss

Returns the loss of the training cycle.

Definition at line 310 of file EventArgs.cs.

◆ SmoothedLoss

double MyCaffe.common.TrainingIterationArgs< T >.SmoothedLoss
get

Retunrs the average loss after the training cycle.

Definition at line 318 of file EventArgs.cs.

◆ WeightsUpdated

bool MyCaffe.common.TrainingIterationArgs< T >.WeightsUpdated
get

Returns whether or not the weights have been updated.

Definition at line 334 of file EventArgs.cs.


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