🚀 wav2vec 2.0 with CTC trained on LibriSpeech
This repository offers all the essential tools for automatic speech recognition. It uses an end - to - end system pretrained on LibriSpeech (English) within SpeechBrain. For an enhanced experience, we recommend learning more about SpeechBrain.
🚀 Quick Start
This repository provides all the necessary tools to perform automatic speech recognition from an end - to - end system pretrained on LibriSpeech (English Language) within SpeechBrain. For a better experience, we encourage you to learn more about SpeechBrain.
The performance of the model is as follows:
Release |
Test clean WER |
Test other WER |
GPUs |
24 - 03 - 22 |
1.90 |
3.96 |
1xA100 40GB |
✨ Features
- Pipeline Composition: This ASR system consists of two different but linked blocks: a unigram tokenizer that converts words into characters (trained with English train transcriptions) and an acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model ([wav2vec2 - large - 960h - lv60 - self](https://huggingface.co/facebook/wav2vec2 - large - 960h - lv60 - self)) is combined with two DNN layers and finetuned on LibriSpeech. The final acoustic representation is fed into the CTC.
- Audio Compatibility: The system is trained with 16kHz single - channel recordings. The code will automatically normalize audio (resampling + mono channel selection) when calling transcribe_file if necessary.
📦 Installation
Install SpeechBrain
First, install transformers
and SpeechBrain
using the following command:
pip install speechbrain transformers
Please note that we encourage you to read our tutorials and learn more about SpeechBrain.
💻 Usage Examples
Transcribing your own audio files (in English)
from speechbrain.inference.ASR import EncoderASR
asr_model = EncoderASR.from_hparams(source="speechbrain/asr - wav2vec2 - librispeech", savedir="pretrained_models/asr - wav2vec2 - librispeech")
asr_model.transcribe_file("speechbrain/asr - wav2vec2 - commonvoice - en/example.wav")
Inference on GPU
To perform inference on the GPU, add run_opts={"device":"cuda"}
when calling the from_hparams
method.
Parallel Inference on a Batch
To figure out how to transcribe a batch of input sentences in parallel using a pre - trained model, see this Colab notebook.
Training
The model was trained with SpeechBrain. To train it from scratch, follow these steps:
- Clone SpeechBrain:
git clone https://github.com/speechbrain/speechbrain/
- Install it:
cd speechbrain
pip install -r requirements.txt
pip install -e .
- Run Training:
cd recipes/LibriSpeech/ASR/CTC
python train_with_wav2vec.py hparams/train_en_with_wav2vec.yaml --data_folder=your_data_folder
You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1pg0QzW - LqAISG8Viw_lUTGjXwOqh7gkl?usp=sharing).
Limitations
The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.
📚 Documentation
Pipeline description
This ASR system is composed of 2 different but linked blocks:
- Tokenizer (unigram) that transforms words into characters and trained with the train transcriptions (EN).
- Acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model ([wav2vec2 - large - 960h - lv60 - self](https://huggingface.co/facebook/wav2vec2 - large - 960h - lv60 - self)) is combined with two DNN layers and finetuned on LibriSpeech. The obtained final acoustic representation is given to the CTC.
The system is trained with recordings sampled at 16kHz (single channel). The code will automatically normalize your audio (i.e., resampling + mono channel selection) when calling transcribe_file if needed.
📄 License
This project is licensed under the apache - 2.0
license.
About SpeechBrain
- Website: https://speechbrain.github.io/
- Code: https://github.com/speechbrain/speechbrain/
- HuggingFace: https://huggingface.co/speechbrain/
Citing SpeechBrain
Please cite SpeechBrain if you use it for your research or business.
@misc{speechbrain,
title={{SpeechBrain}: A General - Purpose Speech Toolkit},
author={Mirco Ravanelli and Titouan Parcollet and Peter Plantinga and Aku Rouhe and Samuele Cornell and Loren Lugosch and Cem Subakan and Nauman Dawalatabad and Abdelwahab Heba and Jianyuan Zhong and Ju - Chieh Chou and Sung - Lin Yeh and Szu - Wei Fu and Chien - Feng Liao and Elena Rastorgueva and François Grondin and William Aris and Hwidong Na and Yan Gao and Renato De Mori and Yoshua Bengio},
year={2021},
eprint={2106.04624},
archivePrefix={arXiv},
primaryClass={eess.AS},
note={arXiv:2106.04624}
}