waste-sorting

Pseudo-labeling as a boost in waste classification task

Self-training method relies on making inference on unlabeled data and using created pseudo-labels in further training.