Conv1d module with customized initialization.

1x1 Conv1d with customized initialization.

Conv2d module with customized initialization.

CARDINAL类

中文数字/数位字符 每个字符除繁简体外还有一个额外的大写字符 e.g. '陆' 和 '陸'

CausalConv1d module with customized initialization.

Layer normalization module. :param int nout: output dim size :param int dim: dimension to be normalized

This module produces sinusoidal positional embeddings of any length. Padding symbols are ignored.

DIGIT类

Log STFT magnitude loss module.

Residual block module in WaveNet.

Definition of sine generator SineGen(samp_rate, harmonic_num = 0, sine_amp = 0.1, noise_std = 0.003, voiced_threshold = 0

Spectral convergence loss module.

TELEPHONE类

Encoder based on a user-supplied vocabulary (file or list).

CausalConvTranspose1d module with customized initialization.

中文数字字符

CyclicnoiseGen_v1 Cyclic noise with a single parameter of beta. Pytorch v1 implementation assumes f_t is also fixed

Override the forward call in lightning so it goes to training and validation step respectively

Override the forward call in lightning so it goes to training and validation step respectively

DATE类

Duration predictor module. This is a module of duration predictor described in `FastSpeech: Fast, Robust and Controllable Text to Speech`_. Th

FRACTION类

用于中文数字系统的数学符号 (繁/简体), e.g. positive = ['正', '正'] negative = ['负', '負'] point = ['点', '點']

MelGAN discriminator module.

MONEY类

中文数字系统

Parallel WaveGAN Generator module.

PERCENTAGE类

Definition of Pulse train generator There are many ways to implement pulse generator. Here, PulseGen is based on SinGen. For a perfect

Residual stack module introduced in MelGAN.

STFT loss module.

STFT loss module.

Filtering input signal with time invariant filter Note: FIRFilter conducted filtering given fixed FIR weight SignalsConv1d convolves tw

SourceModuleCycNoise_v1 SourceModule(sampling_rate, noise_std=0.003, voiced_threshod=0) sampling_rate: sampling_rate in Hz noise_std: st

SourceModule for hn-nsf SourceModule(sampling_rate, harmonic_num=0, sine_amp=0.1, add_noise_std=0.003, voiced_threshod=0) sa

Stretch2d module.

Tensorflow ReflectionPad1d module.

Upsampling network module.

Encodes each byte to an id. For 8-bit strings only.

Encodes each byte to an id and appends the EOS token.

中文字符 每个字符对应简体和繁体, e.g. 简体 = '负', 繁体 = '負' 转换时可转换为简体或繁体

中文数位字符

1x1 Conv1d with customized initialization.

Convolution + upsampling network module.