Bp Sid10 Xlsr

🚀 sid10-xlsr: Wav2vec 2.0 with Sidney Dataset

This project demonstrates a fine - tuned Wav2vec model for Brazilian Portuguese, leveraging the Sidney dataset. The model is tested against multiple available Brazilian Portuguese datasets in this notebook.

✨ Features

• Fine - tuned Model: A Wav2vec 2.0 model fine - tuned for Brazilian Portuguese.
• Multi - dataset Testing: Tested against various Brazilian Portuguese datasets, including CETUC, Common Voice, LaPS BM, MLS, Multilingual TEDx (Portuguese), SID, and VoxForge.

📦 Installation

%%capture
!pip install torch==1.8.2+cu111 torchvision==0.9.2+cu111 torchaudio===0.8.2 -f https://download.pytorch.org/whl/lts/1.8/torch_lts.html
!pip install datasets
!pip install jiwer
!pip install transformers
!pip install soundfile
!pip install pyctcdecode
!pip install https://github.com/kpu/kenlm/archive/master.zip

💻 Usage Examples

Basic Usage

MODEL_NAME = "lgris/sid10-xlsr" 

Advanced Usage

# Imports and dependencies
import jiwer
import torchaudio
from datasets import load_dataset, load_metric
from transformers import (
    Wav2Vec2ForCTC,
    Wav2Vec2Processor,
)
from pyctcdecode import build_ctcdecoder
import torch
import re
import sys

# Helpers
chars_to_ignore_regex = '[\,\?\.\!\;\:\"]'  # noqa: W605

def map_to_array(batch):
    speech, _ = torchaudio.load(batch["path"])
    batch["speech"] = speech.squeeze(0).numpy() 
    batch["sampling_rate"] = 16_000 
    batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower().replace("’", "'")
    batch["target"] = batch["sentence"]
    return batch

def calc_metrics(truths, hypos):
    wers = []
    mers = []
    wils = []
    for t, h in zip(truths, hypos):
        try:
            wers.append(jiwer.wer(t, h))
            mers.append(jiwer.mer(t, h))
            wils.append(jiwer.wil(t, h))
        except: # Empty string?
            pass
    wer = sum(wers)/len(wers)
    mer = sum(mers)/len(mers)
    wil = sum(wils)/len(wils)
    return wer, mer, wil

def load_data(dataset):
    data_files = {'test': f'{dataset}/test.csv'}
    dataset = load_dataset('csv', data_files=data_files)["test"]
    return dataset.map(map_to_array)

# Model
class STT:

    def __init__(self, 
                 model_name, 
                 device='cuda' if torch.cuda.is_available() else 'cpu', 
                 lm=None):
        self.model_name = model_name
        self.model = Wav2Vec2ForCTC.from_pretrained(model_name).to(device)
        self.processor = Wav2Vec2Processor.from_pretrained(model_name)
        self.vocab_dict = self.processor.tokenizer.get_vocab()
        self.sorted_dict = {
            k.lower(): v for k, v in sorted(self.vocab_dict.items(), 
                                            key=lambda item: item[1])
        }
        self.device = device
        self.lm = lm
        if self.lm:            
            self.lm_decoder = build_ctcdecoder(
                list(self.sorted_dict.keys()),
                self.lm
            )

    def batch_predict(self, batch):
        features = self.processor(batch["speech"], 
                                  sampling_rate=batch["sampling_rate"][0], 
                                  padding=True, 
                                  return_tensors="pt")
        input_values = features.input_values.to(self.device)
        attention_mask = features.attention_mask.to(self.device)
        with torch.no_grad():
            logits = self.model(input_values, attention_mask=attention_mask).logits
        if self.lm:
            logits = logits.cpu().numpy()
            batch["predicted"] = []
            for sample_logits in logits:
                batch["predicted"].append(self.lm_decoder.decode(sample_logits))
        else:
            pred_ids = torch.argmax(logits, dim=-1)
            batch["predicted"] = self.processor.batch_decode(pred_ids)
        return batch

# Download datasets
%%capture
!gdown --id 1HFECzIizf-bmkQRLiQD0QVqcGtOG5upI
!mkdir bp_dataset
!unzip bp_dataset -d bp_dataset/

# Tests 
stt = STT(MODEL_NAME)

# CETUC
ds = load_data('cetuc_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("CETUC WER:", wer)

# Common Voice
ds = load_data('commonvoice_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("CV WER:", wer)

# LaPS
ds = load_data('lapsbm_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("Laps WER:", wer)

# MLS
ds = load_data('mls_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("MLS WER:", wer)

# SID
ds = load_data('sid_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("Sid WER:", wer)

# TEDx
ds = load_data('tedx_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("TEDx WER:", wer)

# VoxForge
ds = load_data('voxforge_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("VoxForge WER:", wer)

# Tests with LM
# !find -type f -name "*.wav" -delete
!rm -rf ~/.cache
!gdown --id 1GJIKseP5ZkTbllQVgOL98R4yYAcIySFP  # trained with wikipedia
stt = STT(MODEL_NAME, lm='pt-BR-wiki.word.4-gram.arpa')
# !gdown --id 1dLFldy7eguPtyJj5OAlI4Emnx0BpFywg  # trained with bp
# stt = STT(MODEL_NAME, lm='pt-BR.word.4-gram.arpa')

# CETUC with LM
ds = load_data('cetuc_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("CETUC WER:", wer)

# Common Voice with LM
ds = load_data('commonvoice_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("CV WER:", wer)

# LaPS with LM
ds = load_data('lapsbm_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("Laps WER:", wer)

# MLS with LM
ds = load_data('mls_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("MLS WER:", wer)

# SID with LM
ds = load_data('sid_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("Sid WER:", wer)

# TEDx with LM
ds = load_data('tedx_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("TEDx WER:", wer)

# VoxForge with LM
ds = load_data('voxforge_dataset')
result = ds.map(stt.batch_predict, batched=True, batch_size=8) 
wer, mer, wil = calc_metrics(result["sentence"], result["predicted"])
print("VoxForge WER:", wer)

📚 Documentation

Dataset Information

Summary of Results

📄 License

This project is licensed under the Apache - 2.0 license.