语言: eu
数据集:

• common_voice
  标签:
• audio
• automatic-speech-recognition
• speech
• xlsr-fine-tuning-week
  许可证: apache-2.0
  模型索引:
• 名称: XLSR Wav2Vec2 Euskera Manuel Romero
  结果:
  • 任务:
    名称: 语音识别
    类型: automatic-speech-recognition
    数据集:
    名称: Common Voice eu
    类型: common_voice
    参数: eu
    指标:
    • 名称: 测试WER
      类型: wer
      值: 24.03

Wav2Vec2-Large-XLSR-53-euskera

基于facebook/wav2vec2-large-xlsr-53在巴斯克语（Euskera）上使用Common Voice进行微调。
使用此模型时，请确保语音输入采样率为16kHz。

使用方法

该模型可以直接使用（无需语言模型）如下：

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

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

processor = Wav2Vec2Processor.from_pretrained("mrm8488/wav2vec2-large-xlsr-53-euskera")  
model = Wav2Vec2ForCTC.from_pretrained("mrm8488/wav2vec2-large-xlsr-53-euskera")  

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

# 数据预处理  
# 需要将音频文件读取为数组  
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("预测结果:", processor.batch_decode(predicted_ids))  
print("参考文本:", test_dataset["sentence"][:2])  

评估

可以在Common Voice的巴斯克语测试数据上评估该模型如下：

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

test_dataset = load_dataset("common_voice", "eu", split="test")  
wer = load_metric("wer")  

processor = Wav2Vec2Processor.from_pretrained("mrm8488/wav2vec2-large-xlsr-53-euskera")  
model = Wav2Vec2ForCTC.from_pretrained("mrm8488/wav2vec2-large-xlsr-53-euskera")  
model.to("cuda")  

chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“\\%\\‘\\”\\�]'  
resampler = torchaudio.transforms.Resample(48_000, 16_000)  

# 数据预处理  
# 需要将音频文件读取为数组  
def speech_file_to_array_fn(batch):  
  batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()  
  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)  

# 评估函数  
def evaluate(batch):  
  inputs = processor(batch["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)  
  batch["pred_strings"] = processor.batch_decode(pred_ids)  
  return batch  

result = test_dataset.map(evaluate, batched=True, batch_size=8)  

print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))  

测试结果: 24.03 %

训练

使用了Common Voice的train和validation数据集进行训练。

训练脚本可在???找到