README.md

metadata

license: mit
language: fr
datasets:
  - mozilla-foundation/common_voice_13_0
metrics:
  - per
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - phonemize
model-index:
  - name: Wav2Vec2-base French finetuned for phonemes by LMSSC
    results:
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice v13
          type: mozilla-foundation/common_voice_13_0
          args: fr
        metrics:
          - name: Test PER on Common Voice FR 13.0 | Trained
            type: per
            value: 5.52
          - name: Test PER on Multilingual Librispeech FR | Trained
            type: per
            value: 4.36
          - name: Val PER on Common Voice FR 13.0 | Trained
            type: per
            value: 4.31

Fine-tuned French Voxpopuli v2 wav2vec2-base model for speech-to-phoneme task in French

Fine-tuned facebook/wav2vec2-base-fr-voxpopuli-v2 for French speech-to-phoneme (without language model) using the train and validation splits of Common Voice v13.

Audio samplerate for usage

When using this model, make sure that your speech input is sampled at 16kHz.

Output

As this model is specifically trained for a speech-to-phoneme task, the output is sequence of IPA-encoded words, without punctuation. If you don't read the phonetic alphabet fluently, you can use this excellent IPA reader website to convert the transcript back to audio synthetic speech in order to check the quality of the phonetic transcription.

Training procedure

The model has been finetuned on Coommonvoice-v13 (FR) for 14 epochs on 4x2080 Ti GPUs using a ddp strategy and gradient-accumulation procedure (256 audios per update, corresponding roughly to 25 minutes of speech per update -> 2k updates per epoch)

• Learning rate schedule : Double Tri-state schedule

  • Warmup from 1e-5 for 7% of total updates
  • Constant at 1e-4 for 28% of total updates
  • Linear decrease to 1e-6 for 36% of total updates
  • Second warmup boost to 3e-5 for 3% of total updates
  • Constant at 3e-5 for 12% of total updates
  • Linear decrease to 1e-7 for remaining 14% of updates

• The set of hyperparameters used for training are the same as those detailed in Annex B and Table 6 of wav2vec2 paper.

Usage (with HuggingSound)

The model can be used directly using the HuggingSound library:

import pandas as pd
from huggingsound import SpeechRecognitionModel

model = SpeechRecognitionModel("Cnam-LMSSC/wav2vec2-french-phonemizer")

audio_paths = ["./test_relecture_texte.wav", "./10179_11051_000021.flac"]

# No need for the Audio files to be sampled at 16 kHz here, they are automatically resampled by Huggingsound

transcriptions = model.transcribe(audio_paths)

# (Optionnal) Display results in a table :

df = pd.DataFrame(transcriptions)
df['Audio file'] = pd.DataFrame(audio_paths)
df.set_index('Audio file', inplace=True)

df[['transcription']]

Output :

	Audio file	transcription
0	./test_relecture_texte.wav	ʃapitʁ di də abɛse pəti kɔ̃t də ʒyl ləmɛtʁ ɑ̃ʁʒistʁe puʁ libʁivɔksɔʁɡ ibis dɑ̃ la bas kuʁ dœ̃ ʃato sə tʁuva paʁmi tut sɔʁt də volaj œ̃n ibis ʁɔz
1	./10179_11051_000021.flac	kɛl dɔmaʒ kə sə nə swa pa dy sykʁ supiʁa se foʁaz ɑ̃ pasɑ̃ sa lɑ̃ɡ syʁ la vitʁ fɛ̃ dy ʃapitʁ kɛ̃z ɑ̃ʁʒistʁe paʁ sonjɛ̃ sɛt ɑ̃ʁʒistʁəmɑ̃ fɛ paʁti dy domɛn pyblik

Usage