Wav2Vec2-Base-VoxPopuli-Finetuned

在 VoxPopuli corpus 的10K未标记子集上预训练的基础模型，并在en的转录数据上进行微调（有关更多信息，请参阅论文的表1）。

论文： VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretation

作者：Changhan Wang，Morgane Riviere，Ann Lee，Anne Wu，Chaitanya Talnikar，Daniel Haziza，Mary Williamson，Juan Pino，Emmanuel Dupoux来自Facebook AI

有关更多信息，请参见官方网站： here

用于推断的用法

下面展示了如何在 Common Voice dataset 的样本上使用该模型进行推断

#!/usr/bin/env python3
from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
from datasets import load_dataset
import torchaudio
import torch

# resample audio

# load model & processor
model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-10k-voxpopuli-ft-en")
processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-10k-voxpopuli-ft-en")

# load dataset
ds = load_dataset("common_voice", "en", split="validation[:1%]")

# common voice does not match target sampling rate
common_voice_sample_rate = 48000
target_sample_rate = 16000

resampler = torchaudio.transforms.Resample(common_voice_sample_rate, target_sample_rate)


# define mapping fn to read in sound file and resample
def map_to_array(batch):
    speech, _ = torchaudio.load(batch["path"])
    speech = resampler(speech)
    batch["speech"] = speech[0]
    return batch


# load all audio files
ds = ds.map(map_to_array)

# run inference on the first 5 data samples
inputs = processor(ds[:5]["speech"], sampling_rate=target_sample_rate, return_tensors="pt", padding=True)

# inference
logits = model(**inputs).logits
predicted_ids = torch.argmax(logits, axis=-1)

print(processor.batch_decode(predicted_ids))