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.
FashionMNIST - Gray-Scale CIFAR-10 - MNIST
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.
; 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.
; ;\ ; console.logX; // =>/*{ data: <Uint8Array>, shape: [60000, 28, 28], type: 'uint8',}*/
Types
Some of the types that will be returned:
Dataset
// get a dataset object; // access the training tensors// [ featureTensor, targetTensor ]; // access the testing set tensors; // 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; // get the number of classes; // get number of samples in the training set; // get number of samples in the testing set;
DataTensor
// get a dataset object; // get two DataTensor objects; // get the raw data// this will be a flat TypedArray; // get the shape of the tensor // get the type of the tensor// this is redundant with the type of the TypedArray;
@tensorflow/tfjs
;; ;; ; ;
Datasets
Images
MNIST
; ;
@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
; ;
@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
; ;
@techreport{krizhevsky2009learning,
title={Learning multiple layers of features from tiny images},
author={Krizhevsky, Alex and Hinton, Geoffrey},
year={2009},
institution={Citeseer}
}
Audio
Deterding
; ;
@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" }