Wav2vec2 Xls R 1b Portuguese

🚀 Fine-tuned XLS-R 1B model for speech recognition in Portuguese

This fine-tuned model is based on facebook/wav2vec2-xls-r-1b and trained on Portuguese using the train and validation splits of Common Voice 8.0, CORAA, Multilingual TEDx, and Multilingual LibriSpeech. When using this model, ensure 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.

🚀 Quick Start

✨ Features

• Automatic Speech Recognition: Capable of accurately transcribing Portuguese speech.
• Fine-tuned on Multiple Datasets: Utilizes data from Common Voice 8.0, CORAA, Multilingual TEDx, and Multilingual LibriSpeech.

📦 Installation

No specific installation steps are provided in the original README.

💻 Usage Examples

Basic Usage

Using the HuggingSound library:

from huggingsound import SpeechRecognitionModel

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

transcriptions = model.transcribe(audio_paths)

Advanced Usage

Writing your own inference script:

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

LANG_ID = "pt"
MODEL_ID = "jonatasgrosman/wav2vec2-xls-r-1b-portuguese"
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)

📚 Documentation

Evaluation Commands

1. To evaluate on mozilla-foundation/common_voice_8_0 with split test

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

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

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

🔧 Technical Details

📄 License

This model is licensed under the Apache-2.0 license.

📦 Model Index

• Name: XLS-R Wav2Vec2 Portuguese by Jonatas Grosman
• Results:
  • Task: Automatic Speech Recognition
    • Dataset: Common Voice 8 (mozilla-foundation/common_voice_8_0 with pt args)
      • Metrics:
        • Test WER: 8.7
        • Test CER: 2.55
        • Test WER (+LM): 6.04
        • Test CER (+LM): 1.98
  • Task: Automatic Speech Recognition
    • Dataset: Robust Speech Event - Dev Data (speech-recognition-community-v2/dev_data with pt args)
      • Metrics:
        • Dev WER: 24.23
        • Dev CER: 11.3
        • Dev WER (+LM): 19.41
        • Dev CER (+LM): 10.19
  • Task: Automatic Speech Recognition
    • Dataset: Robust Speech Event - Test Data (speech-recognition-community-v2/eval_data with pt args)
      • Metrics:
        • Test WER: 18.8

📖 Citation

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

@misc{grosman2021xlsr-1b-portuguese,
  title={Fine-tuned {XLS-R} 1{B} model for speech recognition in {P}ortuguese},
  author={Grosman, Jonatas},
  howpublished={\url{https://huggingface.co/jonatasgrosman/wav2vec2-xls-r-1b-portuguese}},
  year={2022}
}