The d8 Package

d8 is a Python package to allow you to use your own datasets in various machine learning frameworks by a few lines of Python codes. It also provides hundreds of build-in datasets with a great of diversity to test your machine learning algorithms.

Installation

d8 is a light-weight python package. The easiest way to install it is through pip

pip install d8

Datasets

You can check our examples to construct datasets and build-in datasets by selecting the problem type you are interested.

Image Classificationimage_classification/getting_started.html

Datasets to recognize an object in an image.                     

Object Detectionobject_detection/getting_started.html

Datasets to detect multiple objects with their bounding boxes in an image.

Tabular Classificationtabular_classification/getting_started.html

Datasets to predict values in a column based on other columns’ values.

Semantic Segmentationsemantic_segmentation/dataset.html

Datasets to associate each pixel of an image with a categorical label.

Table of Contents

• Image Classification
  • Getting Started
  • Uses cases
  • Constructing Datasets
  • Built-in Datasets
  • The Dataset class
• Object Detection Datasets
  • Getting Started
  • Constructing Datasets
  • Built-in Datasets
  • The dataset class
• Semantic Segmentation
  • The Dataset API
• Tabular Classification
  • Getting Started
  • Constructing Datasets
  • Built-in Datasets
  • The Dataset API
• Core API
  • Downloading Data
  • Reading Data
  • Base Datasets