[Hackathon 7th] 修复 voxceleb 的类型提升问题

[megemini]

PR types

Bug fixes

PR changes

Others

Describe

修复 voxceleb 的类型提升问题。

错误日志

Traceback (most recent call last):
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/exps/ecapa_tdnn/train.py", line 366, in <module>
    main(args, config)
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/exps/ecapa_tdnn/train.py", line 175, in main
    waveforms = waveform_augment(waveforms, augment_pipeline)
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/io/augment.py", line 896, in waveform_augment
    waveforms_aug = aug(waveforms)  # (N, L)
  File "/home/aistudio/.local/lib/python3.8/site-packages/paddle/nn/layer/layers.py", line 1531, in __call__
    return self.forward(*inputs, **kwargs)
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/io/augment.py", line 772, in forward
    waveforms = self.drop_chunk(waveforms, lengths)
  File "/home/aistudio/.local/lib/python3.8/site-packages/paddle/nn/layer/layers.py", line 1531, in __call__
    return self.forward(*inputs, **kwargs)
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/io/augment.py", line 141, in forward
    clean_amplitude = compute_amplitude(waveforms, lengths.unsqueeze(1))
  File "/home/aistudio/PaddleSpeech/paddlespeech/vector/io/signal_processing.py", line 40, in compute_amplitude
    out = wav_sum / lengths
TypeError: (InvalidType) Type promotion only support calculations between floating-point numbers and between complex and real numbers. But got different data type x: float32, y: int64. (at ../paddle/phi/common/type_promotion.h:234)

@zxcd @Liyulingyue @GreatV @enkilee @yinfan98

[paddle-bot]

Thanks for your contribution!

[zxcd]

LGTM