chilge/nemo

1import math2import numpy as np3import torch4from torch import nn5from torch.nn import functional as F6 7def slice_pitch_segments(x, ids_str, segment_size=4):8  ret = torch.zeros_like(x[:, :segment_size])9  for i in range(x.size(0)):10    idx_str = ids_str[i]11    idx_end = idx_str + segment_size12    ret[i] = x[i, idx_str:idx_end]13  return ret14 15def rand_slice_segments_with_pitch(x, pitch, x_lengths=None, segment_size=4):16  b, d, t = x.size()17  if x_lengths is None:18    x_lengths = t19  ids_str_max = x_lengths - segment_size + 120  ids_str = (torch.rand([b]).to(device=x.device) * ids_str_max).to(dtype=torch.long)21  ret = slice_segments(x, ids_str, segment_size)22  ret_pitch = slice_pitch_segments(pitch, ids_str, segment_size)23  return ret, ret_pitch, ids_str24 25def init_weights(m, mean=0.0, std=0.01):26  classname = m.__class__.__name__27  if classname.find("Conv") != -1:28    m.weight.data.normal_(mean, std)29 30 31def get_padding(kernel_size, dilation=1):32  return int((kernel_size*dilation - dilation)/2)33 34 35def convert_pad_shape(pad_shape):36  l = pad_shape[::-1]37  pad_shape = [item for sublist in l for item in sublist]38  return pad_shape39 40 41def intersperse(lst, item):42  result = [item] * (len(lst) * 2 + 1)43  result[1::2] = lst44  return result45 46 47def kl_divergence(m_p, logs_p, m_q, logs_q):48  """KL(P||Q)"""49  kl = (logs_q - logs_p) - 0.550  kl += 0.5 * (torch.exp(2. * logs_p) + ((m_p - m_q)**2)) * torch.exp(-2. * logs_q)51  return kl52 53 54def rand_gumbel(shape):55  """Sample from the Gumbel distribution, protect from overflows."""56  uniform_samples = torch.rand(shape) * 0.99998 + 0.0000157  return -torch.log(-torch.log(uniform_samples))58 59 60def rand_gumbel_like(x):61  g = rand_gumbel(x.size()).to(dtype=x.dtype, device=x.device)62  return g63 64 65def slice_segments(x, ids_str, segment_size=4):66  ret = torch.zeros_like(x[:, :, :segment_size])67  for i in range(x.size(0)):68    idx_str = ids_str[i]69    idx_end = idx_str + segment_size70    ret[i] = x[i, :, idx_str:idx_end]71  return ret72 73 74def rand_slice_segments(x, x_lengths=None, segment_size=4):75  b, d, t = x.size()76  if x_lengths is None:77    x_lengths = t78  ids_str_max = x_lengths - segment_size + 179  ids_str = (torch.rand([b]).to(device=x.device) * ids_str_max).to(dtype=torch.long)80  ret = slice_segments(x, ids_str, segment_size)81  return ret, ids_str82 83 84def rand_spec_segments(x, x_lengths=None, segment_size=4):85  b, d, t = x.size()86  if x_lengths is None:87    x_lengths = t88  ids_str_max = x_lengths - segment_size89  ids_str = (torch.rand([b]).to(device=x.device) * ids_str_max).to(dtype=torch.long)90  ret = slice_segments(x, ids_str, segment_size)91  return ret, ids_str92 93 94def get_timing_signal_1d(95    length, channels, min_timescale=1.0, max_timescale=1.0e4):96  position = torch.arange(length, dtype=torch.float)97  num_timescales = channels // 298  log_timescale_increment = (99      math.log(float(max_timescale) / float(min_timescale)) /100      (num_timescales - 1))101  inv_timescales = min_timescale * torch.exp(102      torch.arange(num_timescales, dtype=torch.float) * -log_timescale_increment)103  scaled_time = position.unsqueeze(0) * inv_timescales.unsqueeze(1)104  signal = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)], 0)105  signal = F.pad(signal, [0, 0, 0, channels % 2])106  signal = signal.view(1, channels, length)107  return signal108 109 110def add_timing_signal_1d(x, min_timescale=1.0, max_timescale=1.0e4):111  b, channels, length = x.size()112  signal = get_timing_signal_1d(length, channels, min_timescale, max_timescale)113  return x + signal.to(dtype=x.dtype, device=x.device)114 115 116def cat_timing_signal_1d(x, min_timescale=1.0, max_timescale=1.0e4, axis=1):117  b, channels, length = x.size()118  signal = get_timing_signal_1d(length, channels, min_timescale, max_timescale)119  return torch.cat([x, signal.to(dtype=x.dtype, device=x.device)], axis)120 121 122def subsequent_mask(length):123  mask = torch.tril(torch.ones(length, length)).unsqueeze(0).unsqueeze(0)124  return mask125 126 127@torch.jit.script128def fused_add_tanh_sigmoid_multiply(input_a, input_b, n_channels):129  n_channels_int = n_channels[0]130  in_act = input_a + input_b131  t_act = torch.tanh(in_act[:, :n_channels_int, :])132  s_act = torch.sigmoid(in_act[:, n_channels_int:, :])133  acts = t_act * s_act134  return acts135 136 137def convert_pad_shape(pad_shape):138  l = pad_shape[::-1]139  pad_shape = [item for sublist in l for item in sublist]140  return pad_shape141 142 143def shift_1d(x):144  x = F.pad(x, convert_pad_shape([[0, 0], [0, 0], [1, 0]]))[:, :, :-1]145  return x146 147 148def sequence_mask(length, max_length=None):149  if max_length is None:150    max_length = length.max()151  x = torch.arange(max_length, dtype=length.dtype, device=length.device)152  return x.unsqueeze(0) < length.unsqueeze(1)153 154 155def generate_path(duration, mask):156  """157  duration: [b, 1, t_x]158  mask: [b, 1, t_y, t_x]159  """160  device = duration.device161  162  b, _, t_y, t_x = mask.shape163  cum_duration = torch.cumsum(duration, -1)164  165  cum_duration_flat = cum_duration.view(b * t_x)166  path = sequence_mask(cum_duration_flat, t_y).to(mask.dtype)167  path = path.view(b, t_x, t_y)168  path = path - F.pad(path, convert_pad_shape([[0, 0], [1, 0], [0, 0]]))[:, :-1]169  path = path.unsqueeze(1).transpose(2,3) * mask170  return path171 172 173def clip_grad_value_(parameters, clip_value, norm_type=2):174  if isinstance(parameters, torch.Tensor):175    parameters = [parameters]176  parameters = list(filter(lambda p: p.grad is not None, parameters))177  norm_type = float(norm_type)178  if clip_value is not None:179    clip_value = float(clip_value)180 181  total_norm = 0182  for p in parameters:183    param_norm = p.grad.data.norm(norm_type)184    total_norm += param_norm.item() ** norm_type185    if clip_value is not None:186      p.grad.data.clamp_(min=-clip_value, max=clip_value)187  total_norm = total_norm ** (1. / norm_type)188  return total_norm189