tugstugi/bert-base-mongolian-cased

This repository contains pre-trained Mongolian BERT models trained by tugstugi, enod and sharavsambuu . The repository is based on the following open source projects: google-research/bert, Huggingface/pytorch-pretrained-BERT and yoheikikuta/bert-japanese. The Mongolian-BASE-Mongolian-Cased repository is open source and is hosted by the Mongolian Wikipedia and the 700 million word Mongolian news data set. The repository has 5x TPUs to test and use . The model is a pre-training version of the BERT-Pretraining Procedure. The model,