mnist

Documentation

Overview

Package MNIST is the 60000 + 10000 handwritten digits dataset

Index

• Constants
• Variables
• func Error() error
• func ShuffleInfer()
• func ShuffleTrain()
• type Input
  • func (i *Input) Feature(n int) uint32
• type SmallInput
  • func (i *SmallInput) Feature(n int) uint32

Constants

const ImgSize = 28

original

const SmallImgSize = 13

downscaled

Variables

var SmallTrainSet, SmallInferSet [][SmallImgSize * SmallImgSize]byte

var TrainLabels, InferLabels []byte

var TrainSet, InferSet [][ImgSize * ImgSize]byte

Functions

func Error

func Error() error

Error reports whether there was any error initializing the dataset

func ShuffleInfer

func ShuffleInfer()

ShuffleInfer shuffles the mnist infer dataset

func ShuffleTrain

func ShuffleTrain()

ShuffleTrain shuffles the mnist train dataset

Types

type Input

type Input [ImgSize * ImgSize]byte

Input is the type of input image

func (*Input) Feature

func (i *Input) Feature(n int) uint32

Feature extracts the n-th feature from input

type SmallInput

type SmallInput [SmallImgSize * SmallImgSize]byte

SmallInput is the type of small input image

func (*SmallInput) Feature

func (i *SmallInput) Feature(n int) uint32

Feature extracts the n-th feature from small input