This repo contains models used in the paper An Efficient Multilingual Language Model Compression through Vocabulary Trimming.