This is a pre-trained classifier (341-806 region, seven-level taxonomy), which was trained on the silva_132_99_16S.fna file and majority_taxonomy_7_levels.txt file. Both of the files can be found, respectively, in SILVA_132_QIIME_release/rep_set/rep_set_16S_only/99/ and SILVA_132_QIIME_release/taxonomy/16S_only/99/ in the Silva_132_release.zip file at https://www.arb-silva.de/download/archive/qiime
Please check the ipython notebook for training details.
The trained V3V4 classifier is available at: https://drive.google.com/file/d/1CvmZ1ONAi1dl2_TnzVm4Hc4gG6eSz7a5/
Update: V3V4 classifier trained with sklearn 0.21.2. That is, it is compatiable with QIIME2-2019.7.
Update: This classifier is also compatiable with QIIME2-2019.10 since the version of scikit-learn (0.21.2) is the same.
https://drive.google.com/open?id=15LDBGnDw638egA9hAdpntssQzEW351bl