If you are following the moving pictures tutorial, pay close attention to the “Blue boxes” sometimes they contain really important information that users overlook some times.
So under the Taxonomic analysis section it points you to the tutorial that covers this topic comprehensively about how to train your own. Can you give that a try and then let us know if you still have any questions.
Taxonomic classifiers perform best when they are trained based on your specific sample preparation and sequencing parameters, including the primers that were used for amplification and the length of your sequence reads. Therefore in general you should follow the instructions in Training feature classifiers with q2-feature-classifier to train your own taxonomic classifiers. We provide some common classifiers on our data resources page, including Silva-based 16S classifiers, though in the future we may stop providing these in favor of having users train their own classifiers which will be most relevant to their sequence data.