Hungarian ModernBERT
We've used the transtokenizer repo to create a mapped ModernBERT-base model to Hungarian. To create the data, we've used the OpenSubtitles corpus to obtain EN-HU parallel sentences to train the mapping.
We've used the transtokenizer repo to create a mapped ModernBERT-base model to Hungarian. To create the data, we've used the OpenSubtitles corpus to obtain EN-HU parallel sentences to train the mapping.