wav2vec2 large xlsr javanese EU AI Act Compliance Profile

Wav2Vec2-Large-XLSR-Javanese

Fine-tuned facebook/wav2vec2-large-xlsr-53 on the OpenSLR High quality TTS data for Javanese. When using this model, make sure that your speech input is sampled at 16kHz.

Usage

The model can be used directly (without a language model) as follows:

import torch
import torchaudio
from datasets import load_dataset, load_metric, Dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
from datasets.utils.download_manager import DownloadManager
from pathlib import Path
import pandas as pd

def load_dataset_javanese():
    urls = [
        "https://www.openslr.org/resources/41/jv_id_female.zip",
        "https://www.openslr.org/resources/41/jv_id_male.zip"
    ]
    dm = DownloadManager()
    download_dirs = dm.download_and_extract(urls)
    data_dirs = [ 
        Path(download_dirs[0])/"jv_id_female/wavs",
        Path(download_dirs[1])/"jv_id_male/wavs",
    ]
    filenames = [ 
        Path(download_dirs[0])/"jv_id_female/line_index.tsv",
        Path(download_dirs[1])/"jv_id_male/line_index.tsv",
    ]
    
    dfs = []
    dfs.append(pd.read_csv(filenames[0], sep='\t', names=["path", "sentence"]))
    dfs.append(pd.read_csv(filenames[1], sep='\t', names=["path", "client_id", "sentence"]))
    dfs[1] = dfs[1].drop(["client_id"], axis=1)
    
    for i, dir in enumerate(data_dirs):
        dfs[i]["path"] = dfs[i].apply(lambda row: str(data_dirs[i]) + "/" + row + ".wav", axis=1)
    df = pd.concat(dfs)
    # df = df.sample(frac=1, random_state=1).reset_index(drop=True)
    dataset = Dataset.from_pandas(df)
    dataset = dataset.remove_columns('__index_level_0__')
    
    return dataset.train_test_split(test_size=0.1, seed=1)

dataset = load_dataset_javanese()
test_dataset = dataset['test']

processor = Wav2Vec2Processor.from_pretrained("cahya/wav2vec2-large-xlsr-javanese")
model = Wav2Vec2ForCTC.from_pretrained("cahya/wav2vec2-larg

Your obligations as a wav2vec2 large xlsr javanese deployer

Track your progress

Sign up to track obligations, receive deadline reminders, and generate audit packages.

Tags

Languages

Library