Package deepnetts.cudnn

Class FullyConnectedForwardCuda

java.lang.Object

deepnetts.cudnn.AbstractForwardCuda

deepnetts.cudnn.FullyConnectedForwardCuda

All Implemented Interfaces:

ForwardAccelerator

public final class FullyConnectedForwardCuda
extends AbstractForwardCuda

Constructor Summary

Constructors

Constructor	Description
FullyConnectedForwardCuda(CudaHandles cudaHandles, FullyConnectedLayer layer)
FullyConnectedForwardCuda(CudaHandles cudaHandles, TensorBase outputs, TensorBase weights, TensorBase biases, TensorBase inputs)

Method Summary

Modifier and Type	Method	Description
void	forward()	Performs forward pass.

Methods inherited from class deepnetts.cudnn.AbstractForwardCuda

copyTensorsToDevice

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FullyConnectedForwardCuda

public FullyConnectedForwardCuda(CudaHandles cudaHandles,
 TensorBase outputs,
 TensorBase weights,
 TensorBase biases,
 TensorBase inputs)

Parameters:

cudaHandles - cuda library context handles

outputs - 1D tensor with outputs for this layer

weights - 2D tensor with weights

biases - 1D tensor with biases

inputs - 1D tensor with inputs from previous layer maybe change order of tensors? i, w, o | mada out = W * X where * is dot product make sure that outputs contains biases and the beta is 1 Ovo izbaci imas novi konstruktor ispod koji je zamena ikoji mora da ima referncu na layer KOristim je samo za testiranje

FullyConnectedForwardCuda

public FullyConnectedForwardCuda(CudaHandles cudaHandles,
 FullyConnectedLayer layer)

Method Details

forward

public void forward()

Description copied from class: AbstractForwardCuda

Performs forward pass.

Specified by:

forward in interface ForwardAccelerator

Specified by:

forward in class AbstractForwardCuda