Wav2vec2 Georgian Daytona

🚀 格魯吉亞語Wav2Vec2-Large-XLSR-53模型

本項目基於格魯吉亞語對 facebook/wav2vec2-large-xlsr-53 模型進行了微調，使用的數據集為 Common Voice。使用此模型時，請確保輸入的語音採樣率為 16kHz。

數據集

• common_voice

評估指標

• wer

標籤

• audio
• automatic-speech-recognition
• speech
• xlsr-fine-tuning-week

許可證

apache-2.0

模型信息

🚀 快速開始

本模型是基於Transformer架構的語音識別模型，通過在格魯吉亞語數據集上微調預訓練模型，實現了對格魯吉亞語語音的準確識別。

✨ 主要特性

• 基於預訓練模型 facebook/wav2vec2-large-xlsr-53 進行微調，在格魯吉亞語語音識別任務上有較好表現。
• 支持直接使用，無需額外的語言模型。

💻 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", "ka", split="test[:2%]") 

processor = Wav2Vec2Processor.from_pretrained("Temur/wav2vec2-Georgian-Daytona") 
model = Wav2Vec2ForCTC.from_pretrained("Temur/wav2vec2-Georgian-Daytona")

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])

📚 詳細文檔

評估

可以按照以下方式在 Common Voice 的格魯吉亞語測試數據上對模型進行評估：

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

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

processor = Wav2Vec2Processor.from_pretrained("Temur/wav2vec2-Georgian-Daytona") 
model = Wav2Vec2ForCTC.from_pretrained("Temur/wav2vec2-Georgian-Daytona")
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：48.34 %

訓練

Common Voice 的 train、validation 等數據集用於模型訓練。訓練腳本可參考 此處。

📄 許可證

本項目採用 apache-2.0 許可證。