metadata

language:
  - uk
tags:
  - text2text-generation
  - punctuation prediction
  - punctuation
library_name: generic
license: mit
metrics:
  - f1
datasets:
  - ubertext2.0
widget:
  - text: доброго вечора ми з україни

Ukrainian model to restore punctuation and capitalization

This is the NeMo model to restore punctuation and capitalization in sentences, trained on 10m+ sentences from UberText 2.0 corpus. Basic transformer under the hood is bert-base-multilingual-cased.

Model restores the following punctuations -- [? . ,].

It also restores capitalization of words.