How to detect and classify multiple sound at the same time? about audioclassification-pytorch HOT 2 CLOSED

您可以对您的音频进行分割，如3秒分割一次，每段音频可以有重复部分，例如：第一段：0-3s，第二段：2-5s.....然后使用这些分割的音频去推理，就可以实现你说的识别多个类别。

machine translation：

You can split your audio, such as 3 seconds split once, each audio can have repeated parts, for example: the first segment: 0-3s, the second segment: 2-5s..... Then use these audio segmentation to reasoning, can achieve what you said to identify multiple categories.

from audioclassification-pytorch.

@iotproductions like this:

predictor = MAClsPredictor(configs=args.configs,
                           model_path=args.model_path,
                           use_gpu=args.use_gpu)

sample, sr = soundfile.read(args.audio_path)

for i in range(0, int(len(sample)/sr), 2):
    data = sample[i*sr:(i + 3)*sr]
    label, score = predictor.predict(audio_data=data, sample_rate=sr)
    print(label)

from audioclassification-pytorch.