wav2vec 2.0 with CTC trained on MEDIA

This repository provides all the necessary tools to perform spoken language understanding from an end-to-end system pretrained on MEDIA (French Language) within SpeechBrain. Its original SpeechBrain recipe follows the paper of G. Laperrière, V. Pelloin, A. Caubriere, S. Mdhaffar, N. Camelin, S. Ghannay, B. Jabaian, Y. Estève, The Spoken Language Understanding MEDIA Benchmark Dataset in the Era of Deep Learning: data updates, training and evaluation tools. Find more about the MEDIA corpus and semantic concepts within the Media ASR (ELRA-S0272) and Media SLU (ELRA-E0024) resources. For a better experience, we encourage you to learn more about SpeechBrain.

The performance of the model is the following:

Release	Test ChER	Test CER	Test CVER	GPUs
22-02-23	7.78	24.88	35.77	1xV100 32GB

Pipeline description

This SLU system is composed of an acoustic model (wav2vec2.0 + CTC). A pretrained wav2vec 2.0 model (LeBenchmark/wav2vec2-FR-3K-large) is combined with three DNN layers and finetuned on MEDIA. The obtained final acoustic representation is given to the CTC greedy decoder.

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.

Install SpeechBrain

First of all, please install tranformers and SpeechBrain with the following command:

pip install speechbrain transformers

Please notice that we encourage you to read our tutorials and learn more about SpeechBrain.

Transcribing and semantically annotating your own audio files (in French)

from speechbrain.inference.ASR import EncoderASR

asr_model = EncoderASR.from_hparams(source="speechbrain/slu-wav2vec2-ctc-MEDIA-full", savedir="pretrained_models/slu-wav2vec2-ctc-MEDIA-full")
asr_model.transcribe_file('speechbrain/slu-wav2vec2-ctc-MEDIA-full/example-fr.wav')

Inference on GPU

To perform inference on the GPU, add run_opts={"device":"cuda"} when calling the from_hparams method.

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 .

Download MEDIA related files:

Modify placeholders in hparams/train_hf_wav2vec_full.yaml:

data_folder = !PLACEHOLDER
channels_path = !PLACEHOLDER
concepts_path = !PLACEHOLDER

Run Training:

cd recipes/MEDIA/SLU/CTC/
python train_hf_wav2vec.py hparams/train_hf_wav2vec_full.yaml

You can find our training results (models, logs, etc) here.

Limitations

The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.

Referencing SpeechBrain

@misc{SB2021,
    author = {Ravanelli, Mirco and Parcollet, Titouan and Rouhe, Aku and Plantinga, Peter and Rastorgueva, Elena and Lugosch, Loren and Dawalatabad, Nauman and Ju-Chieh, Chou and Heba, Abdel and Grondin, Francois and Aris, William and Liao, Chien-Feng and Cornell, Samuele and Yeh, Sung-Lin and Na, Hwidong and Gao, Yan and Fu, Szu-Wei and Subakan, Cem and De Mori, Renato and Bengio, Yoshua },
    title = {SpeechBrain},
    year = {2021},
    publisher = {GitHub},
    journal = {GitHub repository},
    howpublished = {\\\\url{https://github.com/speechbrain/speechbrain}},
  }

About SpeechBrain

SpeechBrain is an open-source and all-in-one speech toolkit. It is designed to be simple, extremely flexible, and user-friendly. Competitive or state-of-the-art performance is obtained in various domains.

Website: https://speechbrain.github.io/

GitHub: https://github.com/speechbrain/speechbrain

speechbrain
/

slu-wav2vec2-ctc-MEDIA-full