Skip to main content
. 2020 Jan 21;5(1):e00774-19. doi: 10.1128/mSystems.00774-19

FIG 2.

FIG 2

The model training and validation data loss plots are shown. The models were trained with the multispecies data sets generated by aligning to the discrete (left) and concatenated (right) databases. The plots show that the models do not indicate overfitting when iterated 5,000 times.