This file contains training histories and R console output of the experiments with the best performing CNN architecture VGG16_1FC 
supplementing the manuscript "Deep learning-based diatom taxonomy on virtual slides" by Michael Kloster, Daniel Langenkmper, Martin Zurowietz, Bnk Beszteri & Tim W. Nattkemper.

Please refer to the original publication for further details and cite the publication if you make use of this data.

This material is licensed under the the Creative Commons license CC-BY-NC: Creative Commons Attribution-NonCommercial 4.0 International
 
