The sequence of execution Active learning with Fast iteration Step 1 (source): get the source features python3 step1_save_feat_source.py cluster the source anchors python3 step1_cluster_anchors_source.py: select active samples with ratio of 0.01 python3 step1_select_active_samples.py: train stage1 model python3 step1_train_active_suponly.py Step 2 (iterate for n times in target): get the target features python3 step2_n_save_feat_target.py cluster the target anchors python3 step2_n_cluster_anchors_source.py: additionally select active samples python3 step2_n_select_active_samples.py: train model python3 step1_train_active_suponly.py