Class Conv2d
java.lang.Object
ai.beestream.jitnn.nn.Module
ai.beestream.jitnn.nn.Conv2d
- All Implemented Interfaces:
AutoCloseable
2-D convolution over NCHW with a square kernel, PyTorch's
nn.Conv2d semantics and
initialisation: weight and bias uniform in [-1/sqrt(fan_in), 1/sqrt(fan_in)), fan_in =
in/groups * k * k.-
Constructor Summary
ConstructorsConstructorDescriptionConv2d(int in, int out, int kernel, int stride, int padding, int dilation, int groups, boolean bias, long seed) A convolution fromintooutchannels ingroupsgroups (both divisible by it), initialised fromseed.Conv2d(int in, int out, int kernel, int stride, int padding, long seed) A convolution fromintooutchannels with akernelxkernelkernel and a bias, no dilation, one group. -
Method Summary
Methods inherited from class Module
buffer, buffers, child, close, isClosed, isTraining, loadStateDict, parameter, parameterCount, parameters, stateDict, train, zeroGrad
-
Constructor Details
-
Conv2d
public Conv2d(int in, int out, int kernel, int stride, int padding, long seed) A convolution fromintooutchannels with akernelxkernelkernel and a bias, no dilation, one group. -
Conv2d
public Conv2d(int in, int out, int kernel, int stride, int padding, int dilation, int groups, boolean bias, long seed) A convolution fromintooutchannels ingroupsgroups (both divisible by it), initialised fromseed.
-
-
Method Details
-
weight
The weight [out, in/groups, kernel, kernel]. -
bias
The bias [out], or null without one. -
forward
-
forwardRelu
-