mlts-experiment-data

Easily download and import common machine learning datasets including MNIST, gray-scale CIFAR10, and FashionMNIST. Seemlessly integrates with @tensorflow/tfjs tensors making it easy to quickly validate models on commonly used datasets.

Images

FashionMNIST - Gray-Scale CIFAR-10 - MNIST

Audio

Deterding

API

Each dataset has a common api:

download

Downloads the data to the specified folder, recursively building the folder path if necessary. This method does not import or load the data, only downloads the data if necessary. If the data has already been downloaded, this method will not re-download.

import { FashionMnist } from 'mlts-experiment-data';
 
await FashionMnist.download('path/to/download/location');

load

Downloads the data (if necessary) and loads it into an appropriate typed array. Returns a Dataset object.

import { Mnist } from 'mlts-experiment-data';
 
const mnist = await Mnist.load('path/to/download/location');\
 
const [ X, Y ] = mnist.train;
 
console.log(X); // =>
/*
{
    data: <Uint8Array>,
    shape: [60000, 28, 28],
    type: 'uint8',
}
*/

Types

Some of the types that will be returned:

Dataset

// get a dataset object
const data = await Mnist.load();
 
// access the training tensors
// [ featureTensor, targetTensor ]
const [ X, Y ] = data.train;
 
// access the testing set tensors
const [ X_test, Y_test ] = data.test;
 
// get the number of features
// in the case of a multidimensional tensor of rank > 2,
// this is the product of each feature shape.
// For instance MNIST has shape [60000, 28, 28]
// therefore 28 * 28 = 784 features
const features: number = data.features;
 
// get the number of classes
const classes: number = data.classes;
 
// get number of samples in the training set
const samples: number = data.samples;
 
// get number of samples in the testing set
const testSamples: number = data.testSamples;

DataTensor

// get a dataset object
const data = await Mnist.load();
 
// get two DataTensor objects
const [ X, Y ] = data.train;
 
// get the raw data
// this will be a flat TypedArray
const data: Uint8Array | Float32Array | Int32Array = X.data;
 
// get the shape of the tensor
const shape: number[] = X.shape
 
// get the type of the tensor
// this is redundant with the type of the TypedArray
const type: 'float32' | 'uint8' | 'int32' = X.type;

@tensorflow/tfjs

import * as tf from '@tensoflow/tfjs';
import { Mnist } from 'mlts-experiment-data';
 
const data = await Mnist.load();
const [ X, Y ] = data.train;
 
const X_tensor = tf.tensor(X.data, X.shape);
 
const backAgain = await X_tensor.data().then(raw => ({ shape: X_tensor.shape, data: raw }));

Datasets

Images

MNIST

import { Mnist } from 'mlts-experiment-data';
 
const mnist = await Mnist.load();

@article{lecun1998gradient,
  title={Gradient-based learning applied to document recognition},
  author={LeCun, Yann and Bottou, L{\'e}on and Bengio, Yoshua and Haffner, Patrick},
  journal={Proceedings of the IEEE},
  volume={86},
  number={11},
  pages={2278--2324},
  year={1998},
  publisher={IEEE}
}

FashionMNIST

import { FashionMnist } from 'mlts-experiment-data';
 
const fashion = await FashionMnist.load();

@online{xiao2017/online,
  author       = {Han Xiao and Kashif Rasul and Roland Vollgraf},
  title        = {Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms},
  date         = {2017-08-28},
  year         = {2017},
  eprintclass  = {cs.LG},
  eprinttype   = {arXiv},
  eprint       = {cs.LG/1708.07747},
}

Grey-Scale CIFAR-10

import { GrayCifar10 } from 'mlts-experiment-data';
 
const gray_cifar = await GrayCifar10.load();

@techreport{krizhevsky2009learning,
  title={Learning multiple layers of features from tiny images},
  author={Krizhevsky, Alex and Hinton, Geoffrey},
  year={2009},
  institution={Citeseer}
}

Audio

Deterding

import { Deterding } from 'mlts-experiment-data';
 
const vowels = await Deterding.load();

@misc{Dua:2017 ,
author = "Dheeru, Dua and Karra Taniskidou, Efi",
year = "2017",
title = "{UCI} Machine Learning Repository",
url = "http://archive.ics.uci.edu/ml",
institution = "University of California, Irvine, School of Information and Computer Sciences" }