标签:

• 音频
• 自动语音识别
• 音频分类

使用Wav2Vec 2.0进行进食声音分类

使用方法

环境要求

# 所需安装包
!pip install git+https://github.com/huggingface/datasets.git
!pip install git+https://github.com/huggingface/transformers.git
!pip install torchaudio
!pip install librosa

预测示例

import torch
import torch.nn as nn
import torch.nn.functional as F
import torchaudio
from transformers import AutoConfig, Wav2Vec2FeatureExtractor

import librosa
import IPython.display as ipd
import numpy as np
import pandas as pd

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model_name_or_path = "m3hrdadfi/wav2vec2-base-100k-eating-sound-collection"
config = AutoConfig.from_pretrained(model_name_or_path)
feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained(model_name_or_path)
sampling_rate = feature_extractor.sampling_rate
model = Wav2Vec2ForSpeechClassification.from_pretrained(model_name_or_path).to(device)

def speech_file_to_array_fn(path, sampling_rate):
    speech_array, _sampling_rate = torchaudio.load(path)
    resampler = torchaudio.transforms.Resample(_sampling_rate)
    speech = resampler(speech_array).squeeze().numpy()
    return speech


def predict(path, sampling_rate):
    speech = speech_file_to_array_fn(path, sampling_rate)
    inputs = feature_extractor(speech, sampling_rate=sampling_rate, return_tensors="pt", padding=True)
    inputs = {key: inputs[key].to(device) for key in inputs}

    with torch.no_grad():
        logits = model(**inputs).logits

    scores = F.softmax(logits, dim=1).detach().cpu().numpy()[0]
    outputs = [{"Label": config.id2label[i], "Score": f"{round(score * 100, 3):.1f}%"} for i, score in enumerate(scores)]
    return outputs

path = "clips_rd/gummies/gummies_6_04.wav"
outputs = predict(path, sampling_rate)

[
{'Label': '芦荟', 'Score': '0.0%'},
{'Label': '汉堡', 'Score': '0.0%'},
{'Label': '卷心菜', 'Score': '0.0%'},
{'Label': '蜜饯', 'Score': '0.0%'},
{'Label': '胡萝卜', 'Score': '0.0%'},
{'Label': '薯片', 'Score': '0.0%'},
{'Label': '巧克力', 'Score': '0.0%'},
{'Label': '饮料', 'Score': '0.0%'},
{'Label': '薯条', 'Score': '0.0%'},
{'Label': '葡萄', 'Score': '0.0%'},
{'Label': '软糖', 'Score': '99.8%'},
{'Label': '冰淇淋', 'Score': '0.0%'},
{'Label': '果冻', 'Score': '0.1%'},
{'Label': '面条', 'Score': '0.0%'},
{'Label': '泡菜', 'Score': '0.0%'},
{'Label': '披萨', 'Score': '0.0%'},
{'Label': '肋排', 'Score': '0.0%'},
{'Label': '三文鱼', 'Score': '0.0%'},
{'Label': '汤', 'Score': '0.0%'},
{'Label': '鸡翅', 'Score': '0.0%'}
]

评估结果

下表展示了模型整体及各类别的评估分数。

问题反馈?

请通过此处链接提交Github issue。