metadata

language:
  - fr
license: apache-2.0
tags:
  - automatic-speech-recognition
  - mozilla-foundation/common_voice_8_0
  - fr
  - robust-speech-event
datasets:
  - mozilla-foundation/common_voice_8_0
model-index:
  - name: XLS-R Wav2Vec2 French by Jonatas Grosman
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 8
          type: mozilla-foundation/common_voice_8_0
          args: fr
        metrics:
          - name: Test WER
            type: wer
            value: 16.85
          - name: Test CER
            type: cer
            value: 4.66
          - name: Test WER (+LM)
            type: wer
            value: 16.32
          - name: Test CER (+LM)
            type: cer
            value: 4.21
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Robust Speech Event - Dev Data
          type: speech-recognition-community-v2/dev_data
          args: fr
        metrics:
          - name: Dev WER
            type: wer
            value: 22.34
          - name: Dev CER
            type: cer
            value: 9.88
          - name: Dev WER (+LM)
            type: wer
            value: 17.16
          - name: Dev CER (+LM)
            type: cer
            value: 9.38

XLS-R-1B-FRENCH

Fine-tuned facebook/wav2vec2-xls-r-1b on French using the Common Voice 8. When using this model, make sure that your speech input is sampled at 16kHz.

This model has been fine-tuned by the HuggingSound tool, and thanks to the GPU credits generously given by the OVHcloud :)

Usage

Using the HuggingSound library:

from huggingsound import SpeechRecognitionModel

model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-xls-r-1b-french")
audio_paths = ["/path/to/file.mp3", "/path/to/another_file.wav"]

transcriptions = model.transcribe(audio_paths)

Writing your own inference script:

import torch
import librosa
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor

LANG_ID = "fr"
MODEL_ID = "jonatasgrosman/wav2vec2-xls-r-1b-french"
SAMPLES = 10

test_dataset = load_dataset("common_voice", LANG_ID, split=f"test[:{SAMPLES}]")

processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = librosa.load(batch["path"], sr=16_000)
    batch["speech"] = speech_array
    batch["sentence"] = batch["sentence"].upper()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)

with torch.no_grad():
    logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits

predicted_ids = torch.argmax(logits, dim=-1)
predicted_sentences = processor.batch_decode(predicted_ids)

Evaluation Commands

To evaluate on mozilla-foundation/common_voice_8_0 with split test

python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-french --dataset mozilla-foundation/common_voice_8_0 --config fr --split test

To evaluate on speech-recognition-community-v2/dev_data

python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-french --dataset speech-recognition-community-v2/dev_data --config fr --split validation --chunk_length_s 5.0 --stride_length_s 1.0

Citation

If you want to cite this model you can use this:

@misc{grosman2022wav2vec2-xls-r-1b-french,
  title={XLS-R Wav2Vec2 French by Jonatas Grosman},
  author={Grosman, Jonatas},
  publisher={Hugging Face},
  journal={Hugging Face Hub},
  howpublished={\url{https://huggingface.co/jonatasgrosman/wav2vec2-xls-r-1b-french}},
  year={2022}
}