Wav2vec2 Large Xlsr Welsh

🚀 Wav2Vec2-Large-XLSR-Welsh

該項目是在威爾士語Common Voice數據集上對facebook/wav2vec2-large-xlsr-53進行微調得到的模型。使用標準的數據增強方法對數據進行了增強。使用此模型時，請確保語音輸入的採樣率為16kHz，測試結果顯示字錯率（WER）為29.4%。

✨ 主要特性

• 基於facebook/wav2vec2-large-xlsr-53在威爾士語Common Voice數據集上微調。
• 使用標準增強方法對數據進行增強。
• 測試字錯率（WER）為29.4%。

📦 安裝指南

文檔未提及安裝步驟，可參考transformers、torch、torchaudio、datasets庫的官方安裝文檔進行安裝。

💻 使用示例

基礎用法

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

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

processor = Wav2Vec2Processor.from_pretrained("Srulikbdd/Wav2vec2-large-xlsr-welsh")
model = Wav2Vec2ForCTC.from_pretrained("Srulikbdd/Wav2vec2-large-xlsr-welsh")

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

# Preprocessing the datasets.
# We need to read the aduio 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])

高級用法

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

test_dataset = load_dataset("common_voice", "cy", split="test") 

wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("Srulikbdd/Wav2Vec2-large-xlsr-welsh")
model = Wav2Vec2ForCTC.from_pretrained("Srulikbdd/Wav2Vec2-large-xlsr-welsh")

model.to("cuda")

chars_to_ignore_regex = '[\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\,\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\?\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\!\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\-\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\u2013\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\u2014\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\;\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\:\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\%\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\]' 

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

# Preprocessing the datasets.
# We need to read the aduio files as arrays
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)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
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"])))

📚 詳細文檔

模型信息

注意事項

⚠️ 重要提示

使用此模型時，確保語音輸入的採樣率為16kHz。

📄 許可證

本項目採用Apache-2.0許可證。