metadata

language:
  - en
thumbnail: null
tags:
  - automatic-speech-recognition
  - Transducer
  - Conformer
  - pytorch
  - speechbrain
license: apache-2.0
datasets:
  - librispeech
metrics:
  - wer
model-index:
  - name: Conformer-Transducer by SpeechBrain
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: LibriSpeech (clean)
          type: librispeech_asr
          config: clean
          split: test
          args:
            language: en
        metrics:
          - name: Test WER (non-streaming greedy)
            type: wer
            value: 2.72
          - name: Test WER (960ms chunk size, 4 left context chunks)
            type: wer
            value: 3.13
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: LibriSpeech (other)
          type: librispeech_asr
          config: other
          split: test
          args:
            language: en
        metrics:
          - name: Test WER (non-streaming greedy)
            type: wer
            value: 6.47

Conformer for LibriSpeech

This repository provides all the necessary tools to perform automatic speech recognition from an end-to-end system pretrained on LibriSpeech (EN) within SpeechBrain. For a better experience, we encourage you to learn more about SpeechBrain. The performance of the model in full context mode (no streaming) is the following:

Release	Test clean WER	Test other WER	GPUs
24-02-26	2.72	3.13	4xA100 40GB

With streaming, the results with different chunk sizes on test-clean are the following:

	full	cs=32 (1280ms)	24 (960ms)	16 (640ms)	12 (480ms)	8 (320ms)
full	2.72%	-	-	-	-	-
lc=32	-	3.09%	3.07%	3.26%	3.31%	3.44%
16	-	3.10%	3.07%	3.27%	3.32%	3.50%
8	-	3.10%	3.11%	3.31%	3.39%	3.62%
4	-	3.12%	3.13%	3.37%	3.51%	3.80%
2	-	3.19%	3.24%	3.50%	3.79%	4.38%

Pipeline description

TODO

This ASR system is composed of 3 different but linked blocks:

Tokenizer (unigram) that transforms words into subword units and trained with the train transcriptions of LibriSpeech.
Neural language model (Transformer LM) trained on the full 10M words dataset.
Acoustic model made of a conformer encoder and a joint decoder with CTC + transformer. Hence, the decoding also incorporates the CTC probabilities.

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 SpeechBrain with the following command:

pip install speechbrain

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

Transcribing your own audio files (in English)

TODO + streaming TODO

from speechbrain.inference.ASR import EncoderDecoderASR
asr_model = EncoderDecoderASR.from_hparams(source="speechbrain/asr-conformer-transformerlm-librispeech", savedir="pretrained_models/asr-transformer-transformerlm-librispeech")
asr_model.transcribe_file("speechbrain/asr-conformer-transformerlm-librispeech/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

TODO

Please, see this Colab notebook to figure out how to transcribe in parallel a batch of input sentences using a pre-trained model.

Training

TODO

The model was trained with SpeechBrain (Commit hash: 'f73fcc35'). 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/transformer
python train.py hparams/conformer_large.yaml --data_folder=your_data_folder

Limitations

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

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}
}