Armenian-participle-phrase-punctuation

Summary

This repository contains the code and data pipelines for a bachelor thesis project on Armenian NLP, specifically focusing on automatic punctuation of Armenian participle clauses using sequence labeling models. The project employs a knowledge distillation pipeline where a Gemini 2.5 Flash teacher model annotates a large corpus, which is then used to train several student models (BiLSTM, HyeBERT, mBERT, and an ensemble). The work includes comprehensive data processing, model training, and evaluation on human-annotated benchmarks.