Documentation ¶
Index ¶
- Constants
- func Softmax(k tf32.Continuation, node int, a *tf32.V, options ...map[string]interface{}) bool
- func SphericalSoftmax(k tf32.Continuation, node int, a *tf32.V, options ...map[string]interface{}) bool
- type Entropy
- type Network
- func (n *Network) Analyzer(in []iris.Iris)
- func (n *Network) GetEntropy(inputs []iris.Iris) []Entropy
- func (n *Network) GetGradients(inputs []iris.Iris) [][]float32
- func (n *Network) GetVectors(inputs []iris.Iris) []iris.Iris
- func (n *Network) GetVectors2(inputs []iris.Iris) []iris.Iris
- func (n *Network) Iterate(data []float64) float32
Constants ¶
View Source
const ( // B1 exponential decay of the rate for the first moment estimates B1 = 0.9 // B2 exponential decay rate for the second-moment estimates B2 = 0.999 // S is the scaling factor for the softmax S = 1.0 - 1e-300 // Eta is the learning rate Eta = .001 )
View Source
const ( // StateM is the state for the mean StateM = iota // StateV is the state for the variance StateV // StateTotal is the total number of states StateTotal )
Variables ¶
This section is empty.
Functions ¶
func SphericalSoftmax ¶
func SphericalSoftmax(k tf32.Continuation, node int, a *tf32.V, options ...map[string]interface{}) bool
SphericalSoftmax is the spherical softmax function https://arxiv.org/abs/1511.05042
Types ¶
type Entropy ¶
type Entropy struct { Entropy float32 Label string Measures []float64 Index int Order int Optimized float32 }
Entropy is the self entropy of a point
type Network ¶
type Network struct { Rnd *rand.Rand Width int Length int Set tf32.Set Others tf32.Set Input *tf32.V Point *tf32.V L1 tf32.Meta L2 tf32.Meta Cost tf32.Meta I int Points plotter.XYs }
Network is a clustering neural network
func (*Network) GetEntropy ¶
GetEntropy returns the entropy of the network
func (*Network) GetGradients ¶
GetGradients returns the gradients of the network
Click to show internal directories.
Click to hide internal directories.