We provide precomputed features, kernel matrices for three image datasets for attribute-based classification and zero-shot learning.

• Animals with Attributes [1,2] (URL)
• aPascal/aYahoo [3] (URL)
• SUN Attributes [4] (URL)

Please see [2] for a description of the features and kernels as well as baselines.

version 1.0, May 9 2014: 

• complete data+code (all of the below) (4.7GB !)
  
• only code
• data for Animals with Attributes (3.1GB !)
• data for aPascal/aYahoo (785MB)
• data for SUN+code (685MB)

[1] C. H. Lampert, H. Nickisch, and S. Harmeling. "Learning To Detect Unseen Object Classes by Between-Class Attribute Transfer". In CVPR, 2009. (pdf)

[2] C. H. Lampert, H. Nickisch, and S. Harmeling. "Attribute-Based Classification for Zero-Shot Visual Object Categorization" PAMI, 2014.(pdf)

[3] A. Farhadi, I. Endres, D. Hoiem, and D.A. Forsyth, "Describing Objects by their Attributes", CVPR 2009 (pdf)

[4] G. Patterson and J. Hays, "SUN attribute database: Discovering, annotating, and recognizing scene attributes," CVPR, 2012. (pdf)