Wav2vec2 Large Xlsr Gu

🚀 Wav2Vec2-Large-XLSR-53-古吉拉特语

本项目基于 OpenSLR SLR78 数据集，在古吉拉特语上对 facebook/wav2vec2-large-xlsr-53 模型进行了微调。使用此模型时，请确保输入的语音采样率为 16kHz。

🚀 快速开始

本模型可直接使用（无需语言模型），假设您有一个包含古吉拉特语 sentence 和 path 字段的数据集，可按以下步骤操作。

💻 使用示例

基础用法

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

# test_dataset = #TODO: WRITE YOUR CODE TO LOAD THE TEST DATASET. 
# For sample see the Colab link in Training Section.

processor = Wav2Vec2Processor.from_pretrained("gchhablani/wav2vec2-large-xlsr-gu")
model = Wav2Vec2ForCTC.from_pretrained("gchhablani/wav2vec2-large-xlsr-gu")

resampler = torchaudio.transforms.Resample(48_000, 16_000) # The original data was with 48,000 sampling rate. You can change it according to your input.

# Preprocessing the datasets.
# We need to read the audio 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_eval = test_dataset_eval.map(speech_file_to_array_fn)
inputs = processor(test_dataset_eval["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_eval["sentence"][:2])

📚 详细文档

评估

可按以下方式在 OpenSLR 古吉拉特语数据的 10% 上评估该模型：

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

# test_dataset = #TODO: WRITE YOUR CODE TO LOAD THE TEST DATASET. For sample see the Colab link in Training Section.

wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("gchhablani/wav2vec2-large-xlsr-gu")
model = Wav2Vec2ForCTC.from_pretrained("gchhablani/wav2vec2-large-xlsr-gu")
model.to("cuda")

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

# Preprocessing the datasets.
# We need to read the audio 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"])))

测试结果：23.55 %

训练

在去除了一些包含罗马字符的示例后，使用 OpenSLR 古吉拉特语男性 + 女性数据集的 90% 进行训练。 训练使用的 Colab 笔记本可在 此处 找到。

📄 许可证

本项目采用 Apache-2.0 许可证。

📋 模型信息