Published June 1, 2022 | Version 1.0
Dataset Open

EMNIST-DA: A dataset for studying measurement shift

Authors/Creators

  • 1. University of Edinburgh and MPI-IS Tuebingen
  • 2. University of Edinburgh
  • 3. MPI-IS Tuebingen

Description

emnist_large.tar.gz contains the EMNIST-DA dataset consisting of 13 shifted versions of the 47-class extended-MNIST (EMNIST) dataset. As many methods achieve very good performance on MNIST datasets, this dataset was created to be more challenging with 47-classes and some difficult (measurement) shifts.

Source code to generate the dataset is available at https://github.com/cianeastwood/bufr/blob/main/data/emnist.py.

Files

Files (2.8 GB)

Name Size Download all
md5:1aa79e44cd7476a466c4bf855619e17e
2.8 GB Download