chrissewell/whisperX
1
0
Fork 0
mirror of https://github.com/m-bain/whisperX.git synced 2025-07-01 18:17:27 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
m-bain
2022-12-15 01:01:48 +00:00
committed by GitHub
parent 5a9d0fa06b
commit fb2b727e82
1 changed files with 18 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
README.md
View File

@ -1,20 +1,22 @@
<h1 align="center">WhisperX</h1>
<p align="left">Whisper-Based Automatic Speech Recognition with improved timestamp accuracy using forced alignment.
<p align="left">Whisper-Based Automatic Speech Recognition (ASR) with improved timestamp accuracy using forced alignment.
</p>
<h2 align="left">What is it</h2>
<h2 align="left">What is it 🔎</h2>
This repository refines the timestamps of openAI's Whisper model via forced aligment with phoneme-based ASR models (e.g. wav2vec2.0)
**Whisper** is an Automatic Speech Recognition model [developed by OpenAI](https://github.com/openai/whisper), trained on a large dataset of diverse audio. Whilst it does produces highly accurate transcriptions, the corresponding timestamps are at the utterance-level, not per word, and can be inaccurate by several seconds.
**Whisper** is an ASR model [developed by OpenAI](https://github.com/openai/whisper), trained on a large dataset of diverse audio. Whilst it does produces highly accurate transcriptions, the corresponding timestamps are at the utterance-level, not per word, and can be inaccurate by several seconds.
**Phoneme-Based ASR** A suite of models finetuned to recognise the smallest unit of speech distinguishing one word from another, e.g. the element p in "tap". A popular example model is [wav2vec2.0](https://huggingface.co/facebook/wav2vec2-large-960h-lv60-self).
**Forced Alignment** refers to the process by which orthographic transcriptions are aligned to audio recordings to automatically generate phone level segmentation.
<h2 align="left">Setup</h2>
<h2 align="left">Setup ⚙️</h2>
Install this package using
`pip install git+https://github.com/m-bain/whisperx.git`
@ -37,6 +39,7 @@ Using normal whisper out of the box, many transcriptions are out of sync:
https://user-images.githubusercontent.com/36994049/207743923-b4f0d537-29ae-4be2-b404-bb941db73652.mov
Now, using *WhisperX* with forced alignment to wav2vec2.0:
(a) refining segment timestamps
https://user-images.githubusercontent.com/36994049/207744049-5c0ec593-5c68-44de-805b-b1701d6cc968.mov
@ -46,19 +49,24 @@ https://user-images.githubusercontent.com/36994049/207744049-5c0ec593-5c68-44de-
https://user-images.githubusercontent.com/36994049/207744104-ff4faca1-1bb8-41c9-84fe-033f877e5276.mov
<h2 align="left">Limitations</h2>
<h2 align="left">Limitations ⚠️</h2>
- Hacked this up quite quickly, there might be some errors, please raise an issue if you encounter any.
- Currently only working and tested for ENGLISH language.
- Whisper normalises spoken numbers e.g. "fifty seven" to arabic numerals "57". Need to perform this normalization after alignment, so the phonemes can be aligned. Currently just ignores numbers.
- Assumes the initial whisper timestamps are accurate to some degree (within margin of 2 seconds, adjust if needed -- bigger margins more prone to alignment errors)
<h2 align="center">Contact</h2>
<h2 align="left">Coming Soon 🗓</h2>
Contact maxbain[at]robots.ox.ac.uk if you are using this at scale.
[ ] Incorporating word-level speaker diarization
<h2 align="center">Acknowledgements</h2>
[ ] Inference speedup with batch processing
-OpenAI's whisper https://github.com/openai/whisper
<h2 align="left">Contact</h2>
-PyTorch forced alignment tutorial https://pytorch.org/tutorials/intermediate/forced_alignment_with_torchaudio_tutorial.html
Contact maxbain[at]robots.ox.ac.uk non-bug related queries.
<h2 align="left">Acknowledgements 🙏</h2>
Of course, this is mostly just a modification to [openAI's whisper](https://github.com/openai/whisper).
As well as accreditation to this [PyTorch tutorial on forced alignment](https://pytorch.org/tutorials/intermediate/forced_alignment_with_torchaudio_tutorial.html)