Wav2vec2 Large 100k Voxpopuli Ft Common Voice Plus TTS Dataset Portuguese

🚀 Wav2vec2 Large 100k Voxpopuli fine-tuned with Common Voice and TTS-Portuguese Corpus in Portuguese

This model is a fine-tuned version of Wav2vec2 Large 100k Voxpopuli on Portuguese, using the Common Voice 7.0 and TTS-Portuguese Corpus, aiming to solve the problem of automatic speech recognition in Portuguese and provide high - quality speech - to - text conversion.

🚀 Quick Start

Prerequisites

This model is based on the transformers library, and you need to install it first. You can use the following command to install it:

pip install transformers

Loading the Model

from transformers import AutoTokenizer, Wav2Vec2ForCTC
  
tokenizer = AutoTokenizer.from_pretrained("Edresson/wav2vec2-large-100k-voxpopuli-ft-Common-Voice_plus_TTS-Dataset-portuguese")

model = Wav2Vec2ForCTC.from_pretrained("Edresson/wav2vec2-large-100k-voxpopuli-ft-Common-Voice_plus_TTS-Dataset-portuguese")

✨ Features

• Multilingual Adaptability: Based on the pre - trained model of Wav2vec2 Large 100k Voxpopuli, it has strong adaptability to multiple languages.
• High - Quality Fine - Tuning: Fine - tuned with the Common Voice 7.0 and TTS - Portuguese Corpus, it can achieve good performance in Portuguese speech recognition.

📦 Installation

Before using this model, you need to install the transformers library and other necessary dependencies. You can use the following command to install them:

pip install transformers torchaudio datasets jiwer

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, Wav2Vec2ForCTC
  
tokenizer = AutoTokenizer.from_pretrained("Edresson/wav2vec2-large-100k-voxpopuli-ft-Common-Voice_plus_TTS-Dataset-portuguese")
model = Wav2Vec2ForCTC.from_pretrained("Edresson/wav2vec2-large-100k-voxpopuli-ft-Common-Voice_plus_TTS-Dataset-portuguese")

Advanced Usage

Testing with the Common Voice Dataset

dataset = load_dataset("common_voice", "pt", split="test", data_dir="./cv-corpus-6.1-2020-12-11")

resampler = torchaudio.transforms.Resample(orig_freq=48_000, new_freq=16_000)

def map_to_array(batch):
    speech, _ = torchaudio.load(batch["path"])
    batch["speech"] = resampler.forward(speech.squeeze(0)).numpy()
    batch["sampling_rate"] = resampler.new_freq
    batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower().replace("’", "'")
    return batch

ds = dataset.map(map_to_array)
result = ds.map(map_to_pred, batched=True, batch_size=1, remove_columns=list(ds.features.keys()))
print(wer.compute(predictions=result["predicted"], references=result["target"]))

📚 Documentation

Results

For the results check the paper

📄 License

This model is licensed under the apache - 2.0 license.

📋 Model Information