Skip to content

model.md

Latest commit

 

History

History
236 lines (137 loc) · 5.63 KB

model.md

File metadata and controls

236 lines (137 loc) · 5.63 KB
Raw

ReImproveJS > Model

Class: Model

The Model class is handling everything concerning the neural network

Hierarchy

Model

Index

Constructors

Properties

Accessors

Methods

Constructors

constructor

⊕ new Model(config?: SequentialConfig, fitConfig?: ModelFitConfig): Model

Defined in reimprove/model.ts:47

The sequential config is truly optional and is to use only if you want to provide a complete tf.layers implementation of your model. Currently only dense layers are supported but convolutions etc will be implemented quickly. The [[ModelFitConfig]] is concerning the steps, steps per epoch etc ... which is how is the model going to train itself, which is handled by TensorFlowJS.

Parameters:

Param Type Description
Optional config SequentialConfig -
Optional fitConfig ModelFitConfig

Returns: Model

Properties

fitConfig

● fitConfig: ModelFitConfig

Defined in reimprove/model.ts:47

model

● model: Model

Defined in reimprove/model.ts:46

Accessors

InputSize

getInputSize(): number

Defined in reimprove/model.ts:127

Returns: number

OutputSize

getOutputSize(): number

Defined in reimprove/model.ts:123

Returns: number

Methods

addLayer

▸ addLayer(type: LayerType, config: LayerConfig): void

Defined in reimprove/model.ts:71

Method to just add a layer to the model, concatenating it with the previous ones. deprecated: Please now use NeuralNetwork

Parameters:

Param Type Description
type LayerType a type among DENSE, FLATTEN or CONV2D
config LayerConfig -

Returns: void

compile

▸ compile(config: ModelCompileConfig): Model

Defined in reimprove/model.ts:105

To compile the model, refer to [[ModelCompileConfig]] to know exactly what to use, but essentially, give the optimizer ('sgd', 'crossEntropy' , ...) and the loss function ('meanSquaredError', ...), see TFJS's documentation for the exhaustive list.

Parameters:

Param Type Description
config ModelCompileConfig -

Returns: Model

fit

▸ fit(x: Tensor, y: Tensor): Promise<any>

Defined in reimprove/model.ts:114

Parameters:

Param Type
x Tensor
y Tensor

Returns: Promise<any>

loadFromFile

▸ loadFromFile(file: string): Promise<void>

Defined in reimprove/model.ts:61

Parameters:

Param Type
file string

Returns: Promise<void>

predict

▸ predict(x: Tensor, config?: ModelPredictConfig): Result

Defined in reimprove/model.ts:110

Parameters:

Param Type
x Tensor
Optional config ModelPredictConfig

Returns: Result

randomOutput

▸ randomOutput(): number

Defined in reimprove/model.ts:118

Returns: number

<Static> FromNetwork

▸ FromNetwork(network: NeuralNetwork, fitConfig?: ModelFitConfig, name?: string): Model

Defined in reimprove/model.ts:140

Static method to create a Model from a NeuralNetwork. The fit config is optional as well as the name. It returns a prepared model, but not compiled. constructor:

Parameters:

Param Type Default value Description
network NeuralNetwork - -
Optional fitConfig ModelFitConfig - -
Default value name string v4() -

Returns: Model