Skip to main content

A framework-agnostic datasets library for Machine Learning research and education.

Project description

Dataget

Dataget is an easy to use, framework-agnostic, dataset library that gives you quick access to a collection of Machine Learning datasets through a simple API.

Main features:

  • Minimal: Downloads entire datasets with just 1 line of code.
  • Framework Agnostic: Loads data as numpy arrays or pandas dataframes which can be easily used with the majority of Machine Learning frameworks.
  • Transparent: By default stores the data in your current project so you can easily inspect it.
  • Memory Efficient: When a dataset doesn't fit in memory it will return metadata instead so you can iteratively load it.
  • Integrates with Kaggle: Supports loading datasets directly from Kaggle in a variety of formats.

Checkout the documentation for the list of available datasets.

Getting Started

In dataget you just have to do two things:

  • Instantiate a Dataset from our collection.
  • Call the get method to download the data to disk and load it into memory.

Both are usually done in one line:

import dataget


X_train, y_train, X_test, y_test = dataget.image.mnist().get()

This example downloads the MNIST dataset to ./data/image_mnist and loads it as numpy arrays.

Kaggle Support

Kaggle promotes the use of csv files and dataget loves it! With dataget you can quickly download any dataset from the platform and have immediate access to the data:

import dataget

df_train, df_test = dataget.kaggle(dataset="cristiangarcia/pointcloudmnist2d").get(
    files=["train.csv", "test.csv"]
)

To start using Kaggle datasets just make sure you have properly installed and configured the Kaggle API. In the future we want to expand Kaggle support in the following ways:

  • Be able to load any file that numpy or pandas can read.
  • Have generic support for other types of datasets like images, audio, video, etc.
    • e.g dataget.data.kaggle(..., type="image").get(...)

Installation

pip install dataget

Contributing

Adding a new dataset is easy! Read our guide on Creating a Dataset if you are interested in contributing a dataset.

License

MIT License

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for dataget, version 0.4.15
Filename, size File type Python version Upload date Hashes
Filename, size dataget-0.4.15-py3-none-any.whl (16.6 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size dataget-0.4.15.tar.gz (12.4 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page