Wav2vec2 Large Xlsr 53 Latvian

🚀 拉脱维亚语Wav2Vec2-Large-XLSR-53模型

本项目基于Common Voice数据集，对拉脱维亚语数据微调了facebook/wav2vec2-large-xlsr-53模型。使用此模型时，请确保输入的语音采样率为16kHz。

🚀 快速开始

本模型可直接使用（无需语言模型），以下是使用示例：

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

test_dataset = load_dataset("common_voice", "lv", split="test[:2%]")

processor = Wav2Vec2Processor.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")
model = Wav2Vec2ForCTC.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")

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

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = torchaudio.load(batch["path"])
    batch["speech"] = resampler(speech_array).squeeze().numpy()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"][:2], 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)

print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset["sentence"][:2])

✨ 主要特性

• 模型类型：微调后的Wav2Vec2-Large-XLSR-53模型，适用于拉脱维亚语的自动语音识别任务。
• 训练数据：使用Common Voice数据集的拉脱维亚语部分进行训练。
• 评估指标：使用词错误率（WER）进行评估，测试集上的WER为26.89%。

💻 使用示例

基础用法

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

test_dataset = load_dataset("common_voice", "lv", split="test[:2%]")

processor = Wav2Vec2Processor.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")
model = Wav2Vec2ForCTC.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")

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

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = torchaudio.load(batch["path"])
    batch["speech"] = resampler(speech_array).squeeze().numpy()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"][:2], 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)

print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset["sentence"][:2])

高级用法

import torch
import torchaudio
import urllib.request
import tarfile
import pandas as pd
from tqdm.auto import tqdm
from datasets import load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor

# Download the raw data instead of using HF datasets to save disk space 
data_url = "https://voice-prod-bundler-ee1969a6ce8178826482b88e843c335139bd3fb4.s3.amazonaws.com/cv-corpus-6.1-2020-12-11/lv.tar.gz"
filestream = urllib.request.urlopen(data_url)
data_file = tarfile.open(fileobj=filestream, mode="r|gz")
data_file.extractall()

wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")
model = Wav2Vec2ForCTC.from_pretrained("anton-l/wav2vec2-large-xlsr-53-latvian")
model.to("cuda")

cv_test = pd.read_csv("cv-corpus-6.1-2020-12-11/lv/test.tsv", sep='\t')
clips_path = "cv-corpus-6.1-2020-12-11/lv/clips/"

def clean_sentence(sent):
    sent = sent.lower()
    # replace non-alpha characters with space
    sent = "".join(ch if ch.isalpha() else " " for ch in sent)
    # remove repeated spaces
    sent = " ".join(sent.split())
    return sent

targets = []
preds = []

for i, row in tqdm(cv_test.iterrows(), total=cv_test.shape[0]):
    row["sentence"] = clean_sentence(row["sentence"])
    speech_array, sampling_rate = torchaudio.load(clips_path + row["path"])
    resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
    row["speech"] = resampler(speech_array).squeeze().numpy()

    inputs = processor(row["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)

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

    pred_ids = torch.argmax(logits, dim=-1)

    targets.append(row["sentence"])
    preds.append(processor.batch_decode(pred_ids)[0])

print("WER: {:2f}".format(100 * wer.compute(predictions=preds, references=targets)))

测试结果：26.89 %

🔧 训练细节

训练过程使用了Common Voice数据集的train和validation子集。

📄 许可证

本项目采用Apache 2.0许可证。