jimfhahn/annif059 repository overview

A trained FAST service, the nn-ensemble might require up to 32GB RAM on first load. However, using either of the omikuji projects (bonsai and attention) will be much less RAM intensive (on first load) and should provide decent results.

The training data was sourced from Penn University Libraries ( https://www.library.upenn.edu/⁠) and supplemented with data from the POD project ( https://pod.stanford.edu/⁠).

This service contains information from FAST (Faceted Application of Subject Terminology) Data (https://www.oclc.org/research/areas/data-science/fast/download.html⁠ ) which is made available by OCLC Online Computer Library Center, Inc. under the ODC Attribution License (https://www.oclc.org/research/areas/data-science/fast/odcby.html⁠ )