Class MaxPool2d

Hierarchy (View Summary)

Module
- MaxPool2d

Index

Constructors

Properties

Methods

call eval forward named_parameters parameters register train

Constructors

constructor

new MaxPool2d(
    kernel_size: number | number[],
    stride?: number | number[],
    padding?: number | number[],
): MaxPool2d
Parameters
- kernel_size: number | number[]
- Optionalstride: number | number[]
- padding: number | number[] = 0
Returns MaxPool2d
Overrides Module.constructor
- Defined in nn/module.ts:212

Properties

training

training: boolean = true

Methods

call

call(...args: Tensor[]): Tensor
Entry point for running the module. Equivalent to model(x) in Python. In the future, this is where forward hooks will be triggered. Call forward() directly to bypass hooks.
Parameters
- ...args: Tensor[]
Returns Tensor
Inherited from Module.call
- Defined in nn/base.ts:39

eval

eval(): this
Returns this
Inherited from Module.eval
- Defined in nn/base.ts:51

forward

forward(input: Tensor): Tensor
Parameters
- input: Tensor
Returns Tensor
Overrides Module.forward
- Defined in nn/module.ts:223

named_parameters

named_parameters(prefix?: string): [string, nn.parameter.Parameter][]
Parameters
- prefix: string = ''
Returns [string, nn.parameter.Parameter][]
Inherited from Module.named_parameters
- Defined in nn/base.ts:63

parameters

parameters(): nn.parameter.Parameter[]
Returns nn.parameter.Parameter[]
Inherited from Module.parameters
- Defined in nn/base.ts:55

`Protected`register

register(name: string, value: nn.parameter.Parameter | Module): void
Parameters
- name: string
- value: nn.parameter.Parameter | Module
Returns void
Inherited from Module.register
- Defined in nn/base.ts:24

train

train(mode?: boolean): this
Parameters
- mode: boolean = true
Returns this
Inherited from Module.train
- Defined in nn/base.ts:43