Wav2vec2 Large Xlsr 53 Dhivehi V2

🚀 Wav2Vec2-Large-XLSR-53-Dhivehi

本項目是在迪維希語（Dhivehi）上對 facebook/wav2vec2-large-xlsr-53 模型進行微調得到的。使用該模型時，請確保語音輸入的採樣率為 16kHz。

🚀 快速開始

本模型基於 Common Voice 數據集在迪維希語上對預訓練模型進行微調。使用時需注意語音輸入的採樣率。

✨ 主要特性

• 數據集：使用 Common Voice 數據集進行微調。
• 評估指標：使用字錯率（WER）進行評估。
• 適用場景：適用於音頻、自動語音識別等領域。

💻 使用示例

基礎用法

模型可以直接使用（無需語言模型），示例代碼如下：

import torch
import torchaudio
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
test_dataset = load_dataset("common_voice", "dv", split="test[:2%]")
processor = Wav2Vec2Processor.from_pretrained("shahukareem/wav2vec2-large-xlsr-53-dhivehi-v2")
model = Wav2Vec2ForCTC.from_pretrained("shahukareem/wav2vec2-large-xlsr-53-dhivehi-v2")
# 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

可以在 Common Voice 的迪維希語測試數據上對模型進行評估，示例代碼如下：

import torch
import torchaudio
from datasets import load_dataset, load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import re
test_dataset = load_dataset("common_voice", "dv", split="test")
wer = load_metric("wer")
processor = Wav2Vec2Processor.from_pretrained("shahukareem/wav2vec2-large-xlsr-53-dhivehi-v2")
model = Wav2Vec2ForCTC.from_pretrained("shahukareem/wav2vec2-large-xlsr-53-dhivehi-v2")
model.to("cuda")
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\”\�\،\.\؟\!\'\"\–\’]'
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"])))

📄 許可證

本項目採用 Apache 2.0 許可證。