DeepCL
SNAPSHOT
Deep convolutional neural networks using OpenCL
|
Runs multiple learning epochs using Batcher objects. More...
Inherits NetLearnerBase.
Public Member Functions | |
publicapi | NetLearner (Trainer *trainer, Trainable *net, int Ntrain, float *trainData, int *trainLabels, int Ntest, float *testData, int *testLabels, int batchSize) |
virtual publicapi void | reset () |
virtual publicapi bool | tickBatch () |
virtual publicapi bool | getEpochDone () |
virtual publicapi int | getNextEpoch () |
virtual publicapi int | getNextBatch () |
virtual publicapi int | getNTrain () |
virtual publicapi int | getBatchNumRight () |
virtual publicapi float | getBatchLoss () |
virtual publicapi bool | tickEpoch () |
virtual publicapi void | run () |
virtual publicapi bool | isLearningDone () |
Runs multiple learning epochs using Batcher objects.
Handles learning the neural net, ie running multiple epochs. Uses two Batchers, one for training, one for testing, to learn the epochs.
This class expects the data to be already in memory. If the data is really big, wont fit in memory, you probably want to use something more like NetLearnerOnDemand, which can load in a chunk of data from datafiles at a time
publicapi NetLearner::NetLearner | ( | Trainer * | trainer, |
Trainable * | net, | ||
int | Ntrain, | ||
float * | trainData, | ||
int * | trainLabels, | ||
int | Ntest, | ||
float * | testData, | ||
int * | testLabels, | ||
int | batchSize | ||
) |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |
|
virtual |