Masked Language Modeling

[AmitMY]

Feature request

Support [MASK] token prediction for models such as ModernBERT

Motivation

This repository seems like an extremely fast way to run encoder only language models such as ModernBERT, and might be faster for MLM compared to transformers

Your contribution

I don't think I can assist on this one.