Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/kurianbenoy/whisper_normalizer

A python package for whisper normalizer
https://github.com/kurianbenoy/whisper_normalizer

asr asr-benchmark jupyter-notebook nbdev normalizers openai whisper

Last synced: 1 day ago
JSON representation

A python package for whisper normalizer

Host: GitHub
URL: https://github.com/kurianbenoy/whisper_normalizer
Owner: kurianbenoy
License: mit
Created: 2023-03-21T14:57:13.000Z (almost 2 years ago)
Default Branch: main
Last Pushed: 2024-07-05T01:41:42.000Z (6 months ago)
Last Synced: 2024-10-29T21:05:22.394Z (about 2 months ago)
Topics: asr, asr-benchmark, jupyter-notebook, nbdev, normalizers, openai, whisper
Language: Jupyter Notebook
Homepage: https://kurianbenoy.github.io/whisper_normalizer/
Size: 508 KB
Stars: 44
Watchers: 5
Forks: 7
Open Issues: 7
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.bak
- Funding: .github/FUNDING.yml
- License: LICENSE

Awesome Lists containing this project

README

        # whisper_normalizer

> A python package for text standardisation/normalization. It uses

> normalization algorithm mentioned in OpenAI whisper paper. Using

> Whisper normalization can cause issues in Indic languages and other

> low resource languages when using

> [`BasicTextNormalizer`](https://kurianbenoy.github.io/whisper_normalizer/basic.html#basictextnormalizer).

> So normalization in Indic languages is also implemented in this

> package which was derived from

> [indic-nlp-library](https://github.com/anoopkunchukuttan/indic_nlp_library).

[![Github

license](https://img.shields.io/github/license/kurianbenoy/whisper_normalizer.svg)](https://github.com/kurianbenoy/whisper_normalizer/blob/main/LICENSE)

[![Github

Stars](https://img.shields.io/github/stars/kurianbenoy/whisper_normalizer.svg?colorA=orange&colorB=orange&logo=github)](https://github.com/kurianbenoy/whisper_normalizer/stargazers)

[![PyPI

version](https://img.shields.io/pypi/v/whisper-normalizer.svg?style=flat-square&logo=pypi&logoColor=white)](https://pypi.org/project/whisper-normalizer/)

This package is a python implementation of the text

standardisation/normalization approach which is being used in OpenAI

whisper. The code was originally being released as open-source in

[Whisper source code](https://github.com/openai/whisper). More details

about the text normalization approach used by whisper can be found on

Appendix Section C pp.21 the paper [Robust Speech Recognition via

Large-Scale Weak Supervision](https://cdn.openai.com/papers/whisper.pdf)

by OpenAI team.

## Installation of package

``` sh

pip install whisper_normalizer

```

or from github repository

``` sh

pip install git+https://github.com/kurianbenoy/whisper_normalizer.git

```

## How to use the package

- I made a video walk through on how to use the `whisper_normalizer`

  python package.

[Colab Notebook Link of walk

through](https://colab.research.google.com/gist/kurianbenoy/7d27d9ec193a4a97ec7821235bddc506/hello-world_whisper_normalizer.ipynb)

[Github Gist Link of walk

through](https://gist.github.com/kurianbenoy/7d27d9ec193a4a97ec7821235bddc506)

[![Hello world to

whisper_normalizer](https://img.youtube.com/vi/c7trf0zul6g/0.jpg)](https://www.youtube.com/watch?v=c7trf0zul6g)

## Why should we normalize/standardize text?

- In ASR systems it’s important to normalize the text to reduce

  unintentional penalties in metrics like WER, CER etc.

- Text normalization/standardization is process of converting texts in

  different styles into a standardized form, which is a best-effort

  attempt to penalize only when a word error is caused by actually

  mistranscribing a word, and not by formatting or punctuation

  differences.(from [Whisper

  paper](https://cdn.openai.com/papers/whisper.pdf))

## Why use this python package?

This package is a python implementation of the text

standardisation/normalization approach which is being used in OpenAI

whisper text normalizer. If you want to use just text normalization

alone, it’s better to use this instead reimplementing the same thing.

OpenAI approach of text normalization is very helpful and is being used

as normalization step when evaluating competitive models like

[AssemblyAI Conformer-1

model](https://www.assemblyai.com/blog/conformer-1/).

## Models evaluated using Whisper normalization

- OpenAI Whisper

- Massively Multilingual Speech (MMS) models by Meta

- Conformer 1 by AssemblyAI

- Conformer 2 by AssemblyAI

## How to use

OpenAI open source approach of text normalization/standardization is

mentioned in detail Appendix Section C pp.21 the paper [Robust Speech

Recognition via Large-Scale Weak

Supervision](https://cdn.openai.com/papers/whisper.pdf).

Whisper Normalizer by default comes with two classes

[`BasicTextNormalizer`](https://kurianbenoy.github.io/whisper_normalizer/basic.html#basictextnormalizer)

and

[`EnglishTextNormalizer`](https://kurianbenoy.github.io/whisper_normalizer/english.html#englishtextnormalizer)

You can use the same thing in this package as follows:

``` python

from whisper_normalizer.english import EnglishTextNormalizer

english_normalizer = EnglishTextNormalizer()

english_normalizer("I'm a little teapot, short and stout. Tip me over and pour me out!")

```

    'i am a little teapot short and stout tip me over and pour me out'

``` python

from whisper_normalizer.basic import BasicTextNormalizer

normalizer = BasicTextNormalizer()

normalizer("I'm a little teapot, short and stout. Tip me over and pour me out!")

```

    'i m a little teapot short and stout tip me over and pour me out '

## Using BasicTextNormalizer in your mother tongue might be a bad idea

Whisper Text Normalizer is not always recommended to be used. [Dr Kavya

Manohar](https://www.linkedin.com/in/kavya-manohar/) has written a

blogpost on why it might be a bad idea on her [blopost titled Indian

Languages and Text Normalization: Part

1](https://kavyamanohar.com/post/indic-normalizer/).

## This model extends Whisper_normalizer to support Indic languages as well.

The logic for normalization in Indic languages is derived from

[indic-nlp-library](https://github.com/anoopkunchukuttan/indic_nlp_library).

The logic for Malayalam normalization is expanded beyond the Indic NLP

library by

[`MalayalamNormalizer`](https://kurianbenoy.github.io/whisper_normalizer/1b.indic_normalizer.html#malayalamnormalizer).

``` python

from whisper_normalizer.indic_normalizer import MalayalamNormalizer

normalizer = MalayalamNormalizer()

normalizer("എന്റെ കമ്പ്യൂട്ടറിനു് എന്റെ ഭാഷ.")

```

    'എന്റെ കമ്പ്യൂട്ടറിന് എന്റെ ഭാഷ.'