# summary
multilingual tokenizer trained on multilingual data by using the SentencePiece library and the BPE algorithm. 

* vocab size: 100k