torchaudio.prototype.models.hifigan_vocoder¶
- torchaudio.prototype.models.hifigan_vocoder(in_channels: int, upsample_rates: Tuple[int, ...], upsample_initial_channel: int, upsample_kernel_sizes: Tuple[int, ...], resblock_kernel_sizes: Tuple[int, ...], resblock_dilation_sizes: Tuple[Tuple[int, ...], ...], resblock_type: int, lrelu_slope: float) HiFiGANVocoder[source]¶
已弃用
警告
此函数已被弃用。它将在 2.9 版本中被移除。此弃用是过渡 TorchAudio 进入维护阶段的大规模重构工作的一部分。有关更多信息，请参阅 https://github.com/pytorch/audio/issues/3902。
构建 HiFi GAN Vocoder [Kong et al., 2020]。
- 参数
in_channels (int) – 参见
HiFiGANVocoder。upsample_rates (tuple of
int) – 参见HiFiGANVocoder。upsample_initial_channel (int) – 参见
HiFiGANVocoder。upsample_kernel_sizes (tuple of
int) – 参见HiFiGANVocoder。resblock_kernel_sizes (tuple of
int) – 参见HiFiGANVocoder。resblock_dilation_sizes (tuple of tuples of
int) – 参见HiFiGANVocoder。resblock_type (int, 1 or 2) – 参见
HiFiGANVocoder。
- 返回
生成的模型。
- 返回类型
