markusiko's picture
Update README.md
721a9ac verified
|
raw
history blame
1.42 kB
metadata
license: mit
language:
  - ru
metrics:
  - seqeval
tags:
  - generated-from-trainer
  - restore_punctuation
widget:
  - text: почему она ушла несмотря на то что ей было хорошо
  - text: привет как дела
  - text: сколько денег нужно чтобы стать счастливым
  - text: это было сильно смело но глупо

ruBert-base for Punctuation Correction

The model is built upon the foundation of ruBert-base and has been fine-tuned to correctly place punctuation marks in Russian sentences (it predicts the mark after each word).

Some additional info about the model:

  • Fine-Tuning Source: The model has undergone fine-tuning using a diverse dataset comprising over 20,000 paragraphs from Russian literary works.

  • Supported Classes: The model is designed to predict classes following specific punctuation marks: ? ! . , : ... and space (as class O).

  • Input Format: To achieve optimal results, input text should be provided without punctuation marks. The model does not process changes in letter case.

Usage Guidelines

To use the model effectively, follow these guidelines:

  1. Input Text: Feed the model with text excluding punctuation marks.

  2. Letter Case: The model does not recognize changes in letter case.

Authors

  • Mark Stolyarov