A POS (Part-of-speech) tagger to grammatically classify tokens (noun, verb, adjective...) in a sentence.
I've used a pre-trained model name ParsBERT, a transformer-based Persian language understanding model, and a monolingual one based on Google's BERT architecture.
I've used Bijankhan Corpus for this task. You can access it here.