Wav2vec2 Xlsr Punjabi

🚀 Wav2Vec2-Large-XLSR-53-旁遮普語

本項目基於 Common Voice 數據集，在旁遮普語上對 facebook/wav2vec2-large-xlsr-53 模型進行了微調。該模型可用於旁遮普語的自動語音識別任務，為語音處理領域提供了有力支持。

🚀 快速開始

使用此模型時，請確保您的語音輸入採樣率為 16kHz。

✨ 主要特性

• 多領域適用：適用於音頻、自動語音識別等多個領域。
• 基於大規模預訓練：以 facebook/wav2vec2-large-xlsr-53 為基礎，在旁遮普語上進行微調，性能更優。

📦 安裝指南

文檔未提及具體安裝步驟，暫不提供。

💻 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", "pa-IN", split="test")

processor = Wav2Vec2Processor.from_pretrained("gagan3012/wav2vec2-xlsr-punjabi") 
model = Wav2Vec2ForCTC.from_pretrained("gagan3012/wav2vec2-xlsr-punjabi") 

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])

結果：

Prediction: ['ਹਵਾ ਲਾਤ ਵਿੱਚ ਪੰਦ ਛੇ ਇਖਲਾਟਕੀ ਮੁਜਰਮ ਸਨ', 'ਮੈ ਇ ਹਾ ਪੈਸੇ ਲੇਹੜ ਨਹੀਂ ਸੀ ਚੌਨਾ'] Reference: ['ਹਵਾਲਾਤ ਵਿੱਚ ਪੰਜ ਛੇ ਇਖ਼ਲਾਕੀ ਮੁਜਰਮ ਸਨ', 'ਮੈਂ ਇਹ ਪੈਸੇ ਲੈਣੇ ਨਹੀਂ ਸੀ ਚਾਹੁੰਦਾ']

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", "pa-IN", split="test") #TODO: replace {lang_id} in your language code here. Make sure the code is one of the *ISO codes* of [this](https://huggingface.co/languages) site.
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("gagan3012/wav2vec2-xlsr-punjabi") 
model = Wav2Vec2ForCTC.from_pretrained("gagan3012/wav2vec2-xlsr-punjabi") 
model.to("cuda")

chars_to_ignore_regex = '[\\\\\\\\\\\\\\\\,\\\\\\\\\\\\\\\\?\\\\\\\\\\\\\\\\.\\\\\\\\\\\\\\\\!\\\\\\\\\\\\\\\\-\\\\\\\\\\\\\\\\;\\\\\\\\\\\\\\\\:\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\“]'  # TODO: adapt this list to include all special characters you removed from the data
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"])))

測試結果：58.05 %

📚 詳細文檔

評估

該模型可以在 Common Voice 的旁遮普語測試數據上按以下方式進行評估。

# 上述高級用法中的評估代碼

訓練

訓練腳本可在 此處 找到。

📄 許可證

本項目採用 Apache-2.0 許可證。

📦 模型信息