This archive provides supplementary information and trained model weights for the paper “PlanktonFlow: Hands-on Deep Learning Classification of Plankton Images for Biologists”. It contains: A supplementary document describing implementation details not included in the main text, such as loss function algorithms, learning rate scheduling, and early stopping strategies and figurer illustrating additional results. The best trained model weights obtained during the experiments. Code to reproduce the figures and results from the paper. These materials are intended for readers who wish to examine the methodological details of the study or to run inference and compare model performance with the reported metrics. They are not required to train new models using the PlanktonFlow pipeline, but they provide a deeper look into the implementation and the final trained model. For training new models or adapting the approach, please refer to the primary dataset and the open-source pipeline available on GitHub.