mldata

A framework to layer a standard interface over datasets commonly used in machine learning.

Goals

• Store datasets on disk in their canonical format, for interoperability
• Transparently download datasets on first use
• Handle common dataset tasks:
  • Splitting datasets
  • Data preprocessing
  • Minibatching