Bert Restore Punctuation

🚀 bert-restore-punctuation

This is a BERT-based model designed to restore punctuation and capitalization in English text. It's fine-tuned on Yelp Reviews, and can be used directly or further fine-tuned for domain-specific tasks.

🚀 Quick Start

Installation

First, install the necessary package:

pip install rpunct

Usage Examples

Basic Usage

from rpunct import RestorePuncts
# The default language is 'english'
rpunct = RestorePuncts()
rpunct.punctuate("""in 2018 cornell researchers built a high-powered detector that in combination with an algorithm-driven process called ptychography set a world record
by tripling the resolution of a state-of-the-art electron microscope as successful as it was that approach had a weakness it only worked with ultrathin samples that were
a few atoms thick anything thicker would cause the electrons to scatter in ways that could not be disentangled now a team again led by david muller the samuel b eckert
professor of engineering has bested its own record by a factor of two with an electron microscope pixel array detector empad that incorporates even more sophisticated
3d reconstruction algorithms the resolution is so fine-tuned the only blurring that remains is the thermal jiggling of the atoms themselves""")
# Outputs the following:
# In 2018, Cornell researchers built a high-powered detector that, in combination with an algorithm-driven process called Ptychography, set a world record by tripling the
# resolution of a state-of-the-art electron microscope. As successful as it was, that approach had a weakness. It only worked with ultrathin samples that were a few atoms
# thick. Anything thicker would cause the electrons to scatter in ways that could not be disentangled. Now, a team again led by David Muller, the Samuel B. 
# Eckert Professor of Engineering, has bested its own record by a factor of two with an Electron microscope pixel array detector empad that incorporates even more
# sophisticated 3d reconstruction algorithms. The resolution is so fine-tuned the only blurring that remains is the thermal jiggling of the atoms themselves.

Note: This model works on arbitrarily large English text and uses GPU if available.

✨ Features

• Punctuation Restoration: Restores punctuations such as !, ?, ., ,, -, :, ;, '.
• Capitalization Restoration: Restores the upper-casing of words.
• General Use: Intended for direct use in general English language punctuation restoration.
• Fine-Tuning: Can be further fine-tuned on domain-specific texts for punctuation restoration tasks.

📦 Training Data

Here is the number of product reviews used for fine-tuning the model:

The best convergence was found around 3 epochs.

📚 Documentation

Accuracy

The fine-tuned model obtained the following accuracy on 45,990 held-out text samples:

Below is a breakdown of the model's performance by each label:

📄 License

This project is licensed under the MIT License.

☕ Contact

Contact Daulet Nurmanbetov for questions, feedback, and/or requests for similar models.