grad

package

v0.0.0-...-94b6d7e Latest Latest Go to latest Published: Mar 5, 2023 License: MIT Imports: 5 Imported by: 0

Details

Valid go.mod file

The Go module system was introduced in Go 1.11 and is the official dependency management solution for Go.
Redistributable license

Redistributable licenses place minimal restrictions on how software can be used, modified, and redistributed.
Tagged version

Modules with tagged versions give importers more predictable builds.
Stable version

When a project reaches major version v1 it is considered stable.
Learn more about best practices

Repository

github.com/davidbyttow/gograd

Links

Open Source Insights

Documentation ¶

Index ¶

func DrawDot(n *Scalar) string
type Layer
- func NewLayer(numIn int, numOut int) *Layer
- func (l *Layer) Act(x []*Scalar) []*Scalar
- func (l *Layer) Parameters() Parameters
type MLP
- func NewMLP(numIn int, numOuts []int) *MLP
- func (p *MLP) Act(x []*Scalar) []*Scalar
- func (p *MLP) Parameters() Parameters
type Neuron
- func NewNeuron(numIn int) *Neuron
- func (n *Neuron) Act(x []*Scalar) *Scalar
- func (n *Neuron) Parameters() Parameters
type Parameters
- func (p Parameters) ZeroGrad()
type Scalar

Constants ¶

This section is empty.

Variables ¶

This section is empty.

Functions ¶

func DrawDot ¶

func DrawDot(n *Scalar) string

Types ¶

type Layer ¶

type Layer struct {
	// contains filtered or unexported fields
}

func NewLayer ¶

func NewLayer(numIn int, numOut int) *Layer

func (*Layer) Act ¶

func (l *Layer) Act(x []*Scalar) []*Scalar

func (*Layer) Parameters ¶

func (l *Layer) Parameters() Parameters

type MLP ¶

type MLP struct {
	// contains filtered or unexported fields
}

MLP is a multilayer perceptron

func NewMLP ¶

func NewMLP(numIn int, numOuts []int) *MLP

func (*MLP) Act ¶

func (p *MLP) Act(x []*Scalar) []*Scalar

func (*MLP) Parameters ¶

func (p *MLP) Parameters() Parameters

type Neuron ¶

type Neuron struct {
	// contains filtered or unexported fields
}

func NewNeuron ¶

func NewNeuron(numIn int) *Neuron

func (*Neuron) Act ¶

func (n *Neuron) Act(x []*Scalar) *Scalar

func (*Neuron) Parameters ¶

func (n *Neuron) Parameters() Parameters

type Parameters ¶

type Parameters []*Scalar

func (Parameters) ZeroGrad ¶

func (p Parameters) ZeroGrad()

type Scalar ¶

type Scalar struct {
	Label string
	// contains filtered or unexported fields
}

Scalar is a node like Tensor except for scalar values only It implements simple backpropagation and gradient computation

func MeanSquaredError ¶

func MeanSquaredError(ygts []*Scalar, ypreds []*Scalar) *Scalar

func Scalars ¶

func Scalars(vals ...float64) []*Scalar

func Val ¶

func Val(data float64) *Scalar

func Var ¶

func Var(data float64, label string) *Scalar

func (*Scalar) Add ¶

func (v *Scalar) Add(other *Scalar) *Scalar

func (*Scalar) Backward ¶

func (v *Scalar) Backward()

Backward calls the backward function for each method in a breadth-first way starting with the given root scalar. This iteratively builds the gradient at each depth because of the chain rule

func (*Scalar) Data ¶

func (v *Scalar) Data() float64

func (*Scalar) Descend ¶

func (v *Scalar) Descend(delta float64)

func (*Scalar) Grad ¶

func (v *Scalar) Grad() float64

func (*Scalar) Mul ¶

func (v *Scalar) Mul(other *Scalar) *Scalar

func (*Scalar) Neg ¶

func (v *Scalar) Neg() *Scalar

func (*Scalar) Pow ¶

func (v *Scalar) Pow(exp float64) *Scalar

func (*Scalar) ReLU ¶

func (v *Scalar) ReLU() *Scalar

func (*Scalar) String ¶

func (v *Scalar) String() string

func (*Scalar) Sub ¶

func (v *Scalar) Sub(other *Scalar) *Scalar

func (*Scalar) Tanh ¶

func (v *Scalar) Tanh() *Scalar

Source Files ¶

View all Source files

?	: This menu
/	: Search site
f or F	: Jump to
y or Y	: Canonical URL