StarMirror
/
MockingBird
mirror of https://github.com/babysor/MockingBird
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge pull request #76 from PeijiYang/main 

Support "hifigan vocoder" for realtime synthesis
pull/75/merge
Vega 2021-09-08 23:06:14 +08:00 committed by GitHub
parent 024d88ae96 b60b75ea89
commit dacedfa9cc
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
8 changed files with 656 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
hifigan/config_16k_.json Normal file
View File

@ -0,0 +1,31 @@
{
"resblock": "1",
"num_gpus": 0,
"batch_size": 16,
"learning_rate": 0.0002,
"adam_b1": 0.8,
"adam_b2": 0.99,
"lr_decay": 0.999,
"seed": 1234,
"upsample_rates": [5,5,4,2],
"upsample_kernel_sizes": [10,10,8,4],
"upsample_initial_channel": 512,
"resblock_kernel_sizes": [3,7,11],
"resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
"segment_size": 6400,
"num_mels": 80,
"num_freq": 1025,
"n_fft": 1024,
"hop_size": 200,
"win_size": 800,
"sampling_rate": 16000,
"fmin": 0,
"fmax": 7600,
"fmax_for_loss": null,
"num_workers": 4
}

15
hifigan/env.py Normal file
View File

@ -0,0 +1,15 @@
import os
import shutil
class AttrDict(dict):
def __init__(self, *args, **kwargs):
super(AttrDict, self).__init__(*args, **kwargs)
self.__dict__ = self
def build_env(config, config_name, path):
t_path = os.path.join(path, config_name)
if config != t_path:
os.makedirs(path, exist_ok=True)
shutil.copyfile(config, os.path.join(path, config_name))

73
hifigan/inference.py Normal file
View File

@ -0,0 +1,73 @@
from __future__ import absolute_import, division, print_function, unicode_literals
import glob
import os
import argparse
import json
import torch
import numpy as np
from scipy.io.wavfile import write
from hifigan.env import AttrDict
from hifigan.meldataset import mel_spectrogram, MAX_WAV_VALUE, load_wav
from hifigan.models import Generator
import soundfile as sf
generator = None # type: Generator
_device = None
def load_checkpoint(filepath, device):
assert os.path.isfile(filepath)
print("Loading '{}'".format(filepath))
checkpoint_dict = torch.load(filepath, map_location=device)
print("Complete.")
return checkpoint_dict
def load_model(weights_fpath, verbose=True):
global generator, _device
if verbose:
print("Building hifigan")
with open("./hifigan/config_16k_.json") as f:
data = f.read()
json_config = json.loads(data)
h = AttrDict(json_config)
torch.manual_seed(h.seed)
if torch.cuda.is_available():
# _model = _model.cuda()
_device = torch.device('cuda')
else:
_device = torch.device('cpu')
generator = Generator(h).to(_device)
state_dict_g = load_checkpoint(
weights_fpath, _device
)
generator.load_state_dict(state_dict_g['generator'])
generator.eval()
generator.remove_weight_norm()
def is_loaded():
return generator is not None
def infer_waveform(mel, progress_callback=None):
if generator is None:
raise Exception("Please load hifi-gan in memory before using it")
mel = torch.FloatTensor(mel).to(_device)
mel = mel.unsqueeze(0)
with torch.no_grad():
y_g_hat = generator(mel)
audio = y_g_hat.squeeze()
audio = audio.cpu().numpy()
return audio

178
hifigan/meldataset.py Normal file
View File

@ -0,0 +1,178 @@
import math
import os
import random
import torch
import torch.utils.data
import numpy as np
from librosa.util import normalize
from scipy.io.wavfile import read
from librosa.filters import mel as librosa_mel_fn
MAX_WAV_VALUE = 32768.0
def load_wav(full_path):
sampling_rate, data = read(full_path)
return data, sampling_rate
def dynamic_range_compression(x, C=1, clip_val=1e-5):
return np.log(np.clip(x, a_min=clip_val, a_max=None) * C)
def dynamic_range_decompression(x, C=1):
return np.exp(x) / C
def dynamic_range_compression_torch(x, C=1, clip_val=1e-5):
return torch.log(torch.clamp(x, min=clip_val) * C)
def dynamic_range_decompression_torch(x, C=1):
return torch.exp(x) / C
def spectral_normalize_torch(magnitudes):
output = dynamic_range_compression_torch(magnitudes)
return output
def spectral_de_normalize_torch(magnitudes):
output = dynamic_range_decompression_torch(magnitudes)
return output
mel_basis = {}
hann_window = {}
def mel_spectrogram(y, n_fft, num_mels, sampling_rate, hop_size, win_size, fmin, fmax, center=False):
if torch.min(y) < -1.:
print('min value is ', torch.min(y))
if torch.max(y) > 1.:
print('max value is ', torch.max(y))
global mel_basis, hann_window
if fmax not in mel_basis:
mel = librosa_mel_fn(sampling_rate, n_fft, num_mels, fmin, fmax)
mel_basis[str(fmax)+'_'+str(y.device)] = torch.from_numpy(mel).float().to(y.device)
hann_window[str(y.device)] = torch.hann_window(win_size).to(y.device)
y = torch.nn.functional.pad(y.unsqueeze(1), (int((n_fft-hop_size)/2), int((n_fft-hop_size)/2)), mode='reflect')
y = y.squeeze(1)
spec = torch.stft(y, n_fft, hop_length=hop_size, win_length=win_size, window=hann_window[str(y.device)],
center=center, pad_mode='reflect', normalized=False, onesided=True)
spec = torch.sqrt(spec.pow(2).sum(-1)+(1e-9))
spec = torch.matmul(mel_basis[str(fmax)+'_'+str(y.device)], spec)
spec = spectral_normalize_torch(spec)
return spec
def get_dataset_filelist(a):
# with open(a.input_training_file, 'r', encoding='utf-8') as fi:
# training_files = [os.path.join(a.input_wavs_dir, x.split('|')[0] + '.wav')
# for x in fi.read().split('\n') if len(x) > 0]
# with open(a.input_validation_file, 'r', encoding='utf-8') as fi:
# validation_files = [os.path.join(a.input_wavs_dir, x.split('|')[0] + '.wav')
# for x in fi.read().split('\n') if len(x) > 0]
files = os.listdir(a.input_wavs_dir)
random.shuffle(files)
files = [os.path.join(a.input_wavs_dir, f) for f in files]
training_files = files[: -500]
validation_files = files[-500: ]
return training_files, validation_files
class MelDataset(torch.utils.data.Dataset):
def __init__(self, training_files, segment_size, n_fft, num_mels,
hop_size, win_size, sampling_rate, fmin, fmax, split=True, shuffle=True, n_cache_reuse=1,
device=None, fmax_loss=None, fine_tuning=False, base_mels_path=None):
self.audio_files = training_files
random.seed(1234)
if shuffle:
random.shuffle(self.audio_files)
self.segment_size = segment_size
self.sampling_rate = sampling_rate
self.split = split
self.n_fft = n_fft
self.num_mels = num_mels
self.hop_size = hop_size
self.win_size = win_size
self.fmin = fmin
self.fmax = fmax
self.fmax_loss = fmax_loss
self.cached_wav = None
self.n_cache_reuse = n_cache_reuse
self._cache_ref_count = 0
self.device = device
self.fine_tuning = fine_tuning
self.base_mels_path = base_mels_path
def __getitem__(self, index):
filename = self.audio_files[index]
if self._cache_ref_count == 0:
# audio, sampling_rate = load_wav(filename)
# audio = audio / MAX_WAV_VALUE
audio = np.load(filename)
if not self.fine_tuning:
audio = normalize(audio) * 0.95
self.cached_wav = audio
# if sampling_rate != self.sampling_rate:
# raise ValueError("{} SR doesn't match target {} SR".format(
# sampling_rate, self.sampling_rate))
self._cache_ref_count = self.n_cache_reuse
else:
audio = self.cached_wav
self._cache_ref_count -= 1
audio = torch.FloatTensor(audio)
audio = audio.unsqueeze(0)
if not self.fine_tuning:
if self.split:
if audio.size(1) >= self.segment_size:
max_audio_start = audio.size(1) - self.segment_size
audio_start = random.randint(0, max_audio_start)
audio = audio[:, audio_start:audio_start+self.segment_size]
else:
audio = torch.nn.functional.pad(audio, (0, self.segment_size - audio.size(1)), 'constant')
mel = mel_spectrogram(audio, self.n_fft, self.num_mels,
self.sampling_rate, self.hop_size, self.win_size, self.fmin, self.fmax,
center=False)
else:
mel_path = os.path.join(self.base_mels_path, "mel" + "-" + filename.split("/")[-1].split("-")[-1])
mel = np.load(mel_path).T
# mel = np.load(
# os.path.join(self.base_mels_path, os.path.splitext(os.path.split(filename)[-1])[0] + '.npy'))
mel = torch.from_numpy(mel)
if len(mel.shape) < 3:
mel = mel.unsqueeze(0)
if self.split:
frames_per_seg = math.ceil(self.segment_size / self.hop_size)
if audio.size(1) >= self.segment_size:
mel_start = random.randint(0, mel.size(2) - frames_per_seg - 1)
mel = mel[:, :, mel_start:mel_start + frames_per_seg]
audio = audio[:, mel_start * self.hop_size:(mel_start + frames_per_seg) * self.hop_size]
else:
mel = torch.nn.functional.pad(mel, (0, frames_per_seg - mel.size(2)), 'constant')
audio = torch.nn.functional.pad(audio, (0, self.segment_size - audio.size(1)), 'constant')
mel_loss = mel_spectrogram(audio, self.n_fft, self.num_mels,
self.sampling_rate, self.hop_size, self.win_size, self.fmin, self.fmax_loss,
center=False)
return (mel.squeeze(), audio.squeeze(0), filename, mel_loss.squeeze())
def __len__(self):
return len(self.audio_files)

286
hifigan/models.py Normal file
View File

@ -0,0 +1,286 @@
import torch
import torch.nn.functional as F
import torch.nn as nn
from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d
from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm
from hifigan.utils import init_weights, get_padding
LRELU_SLOPE = 0.1
class ResBlock1(torch.nn.Module):
def __init__(self, h, channels, kernel_size=3, dilation=(1, 3, 5)):
super(ResBlock1, self).__init__()
self.h = h
self.convs1 = nn.ModuleList([
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0],
padding=get_padding(kernel_size, dilation[0]))),
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1],
padding=get_padding(kernel_size, dilation[1]))),
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[2],
padding=get_padding(kernel_size, dilation[2])))
])
self.convs1.apply(init_weights)
self.convs2 = nn.ModuleList([
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
padding=get_padding(kernel_size, 1))),
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
padding=get_padding(kernel_size, 1))),
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
padding=get_padding(kernel_size, 1)))
])
self.convs2.apply(init_weights)
def forward(self, x):
for c1, c2 in zip(self.convs1, self.convs2):
xt = F.leaky_relu(x, LRELU_SLOPE)
xt = c1(xt)
xt = F.leaky_relu(xt, LRELU_SLOPE)
xt = c2(xt)
x = xt + x
return x
def remove_weight_norm(self):
for l in self.convs1:
remove_weight_norm(l)
for l in self.convs2:
remove_weight_norm(l)
class ResBlock2(torch.nn.Module):
def __init__(self, h, channels, kernel_size=3, dilation=(1, 3)):
super(ResBlock2, self).__init__()
self.h = h
self.convs = nn.ModuleList([
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0],
padding=get_padding(kernel_size, dilation[0]))),
weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1],
padding=get_padding(kernel_size, dilation[1])))
])
self.convs.apply(init_weights)
def forward(self, x):
for c in self.convs:
xt = F.leaky_relu(x, LRELU_SLOPE)
xt = c(xt)
x = xt + x
return x
def remove_weight_norm(self):
for l in self.convs:
remove_weight_norm(l)
class Generator(torch.nn.Module):
def __init__(self, h):
super(Generator, self).__init__()
self.h = h
self.num_kernels = len(h.resblock_kernel_sizes)
self.num_upsamples = len(h.upsample_rates)
self.conv_pre = weight_norm(Conv1d(80, h.upsample_initial_channel, 7, 1, padding=3))
resblock = ResBlock1 if h.resblock == '1' else ResBlock2
self.ups = nn.ModuleList()
for i, (u, k) in enumerate(zip(h.upsample_rates, h.upsample_kernel_sizes)):
# self.ups.append(weight_norm(
# ConvTranspose1d(h.upsample_initial_channel//(2**i), h.upsample_initial_channel//(2**(i+1)),
# k, u, padding=(k-u)//2)))
self.ups.append(weight_norm(ConvTranspose1d(h.upsample_initial_channel//(2**i),
h.upsample_initial_channel//(2**(i+1)),
k, u, padding=(u//2 + u%2), output_padding=u%2)))
self.resblocks = nn.ModuleList()
for i in range(len(self.ups)):
ch = h.upsample_initial_channel//(2**(i+1))
for j, (k, d) in enumerate(zip(h.resblock_kernel_sizes, h.resblock_dilation_sizes)):
self.resblocks.append(resblock(h, ch, k, d))
self.conv_post = weight_norm(Conv1d(ch, 1, 7, 1, padding=3))
self.ups.apply(init_weights)
self.conv_post.apply(init_weights)
def forward(self, x):
x = self.conv_pre(x)
for i in range(self.num_upsamples):
x = F.leaky_relu(x, LRELU_SLOPE)
x = self.ups[i](x)
xs = None
for j in range(self.num_kernels):
if xs is None:
xs = self.resblocks[i*self.num_kernels+j](x)
else:
xs += self.resblocks[i*self.num_kernels+j](x)
x = xs / self.num_kernels
x = F.leaky_relu(x)
x = self.conv_post(x)
x = torch.tanh(x)
return x
def remove_weight_norm(self):
print('Removing weight norm...')
for l in self.ups:
remove_weight_norm(l)
for l in self.resblocks:
l.remove_weight_norm()
remove_weight_norm(self.conv_pre)
remove_weight_norm(self.conv_post)
class DiscriminatorP(torch.nn.Module):
def __init__(self, period, kernel_size=5, stride=3, use_spectral_norm=False):
super(DiscriminatorP, self).__init__()
self.period = period
norm_f = weight_norm if use_spectral_norm == False else spectral_norm
self.convs = nn.ModuleList([
norm_f(Conv2d(1, 32, (kernel_size, 1), (stride, 1), padding=(get_padding(5, 1), 0))),
norm_f(Conv2d(32, 128, (kernel_size, 1), (stride, 1), padding=(get_padding(5, 1), 0))),
norm_f(Conv2d(128, 512, (kernel_size, 1), (stride, 1), padding=(get_padding(5, 1), 0))),
norm_f(Conv2d(512, 1024, (kernel_size, 1), (stride, 1), padding=(get_padding(5, 1), 0))),
norm_f(Conv2d(1024, 1024, (kernel_size, 1), 1, padding=(2, 0))),
])
self.conv_post = norm_f(Conv2d(1024, 1, (3, 1), 1, padding=(1, 0)))
def forward(self, x):
fmap = []
# 1d to 2d
b, c, t = x.shape
if t % self.period != 0: # pad first
n_pad = self.period - (t % self.period)
x = F.pad(x, (0, n_pad), "reflect")
t = t + n_pad
x = x.view(b, c, t // self.period, self.period)
for l in self.convs:
x = l(x)
x = F.leaky_relu(x, LRELU_SLOPE)
fmap.append(x)
x = self.conv_post(x)
fmap.append(x)
x = torch.flatten(x, 1, -1)
return x, fmap
class MultiPeriodDiscriminator(torch.nn.Module):
def __init__(self):
super(MultiPeriodDiscriminator, self).__init__()
self.discriminators = nn.ModuleList([
DiscriminatorP(2),
DiscriminatorP(3),
DiscriminatorP(5),
DiscriminatorP(7),
DiscriminatorP(11),
])
def forward(self, y, y_hat):
y_d_rs = []
y_d_gs = []
fmap_rs = []
fmap_gs = []
for i, d in enumerate(self.discriminators):
y_d_r, fmap_r = d(y)
y_d_g, fmap_g = d(y_hat)
y_d_rs.append(y_d_r)
fmap_rs.append(fmap_r)
y_d_gs.append(y_d_g)
fmap_gs.append(fmap_g)
return y_d_rs, y_d_gs, fmap_rs, fmap_gs
class DiscriminatorS(torch.nn.Module):
def __init__(self, use_spectral_norm=False):
super(DiscriminatorS, self).__init__()
norm_f = weight_norm if use_spectral_norm == False else spectral_norm
self.convs = nn.ModuleList([
norm_f(Conv1d(1, 128, 15, 1, padding=7)),
norm_f(Conv1d(128, 128, 41, 2, groups=4, padding=20)),
norm_f(Conv1d(128, 256, 41, 2, groups=16, padding=20)),
norm_f(Conv1d(256, 512, 41, 4, groups=16, padding=20)),
norm_f(Conv1d(512, 1024, 41, 4, groups=16, padding=20)),
norm_f(Conv1d(1024, 1024, 41, 1, groups=16, padding=20)),
norm_f(Conv1d(1024, 1024, 5, 1, padding=2)),
])
self.conv_post = norm_f(Conv1d(1024, 1, 3, 1, padding=1))
def forward(self, x):
fmap = []
for l in self.convs:
x = l(x)
x = F.leaky_relu(x, LRELU_SLOPE)
fmap.append(x)
x = self.conv_post(x)
fmap.append(x)
x = torch.flatten(x, 1, -1)
return x, fmap
class MultiScaleDiscriminator(torch.nn.Module):
def __init__(self):
super(MultiScaleDiscriminator, self).__init__()
self.discriminators = nn.ModuleList([
DiscriminatorS(use_spectral_norm=True),
DiscriminatorS(),
DiscriminatorS(),
])
self.meanpools = nn.ModuleList([
AvgPool1d(4, 2, padding=2),
AvgPool1d(4, 2, padding=2)
])
def forward(self, y, y_hat):
y_d_rs = []
y_d_gs = []
fmap_rs = []
fmap_gs = []
for i, d in enumerate(self.discriminators):
if i != 0:
y = self.meanpools[i-1](y)
y_hat = self.meanpools[i-1](y_hat)
y_d_r, fmap_r = d(y)
y_d_g, fmap_g = d(y_hat)
y_d_rs.append(y_d_r)
fmap_rs.append(fmap_r)
y_d_gs.append(y_d_g)
fmap_gs.append(fmap_g)
return y_d_rs, y_d_gs, fmap_rs, fmap_gs
def feature_loss(fmap_r, fmap_g):
loss = 0
for dr, dg in zip(fmap_r, fmap_g):
for rl, gl in zip(dr, dg):
loss += torch.mean(torch.abs(rl - gl))
return loss*2
def discriminator_loss(disc_real_outputs, disc_generated_outputs):
loss = 0
r_losses = []
g_losses = []
for dr, dg in zip(disc_real_outputs, disc_generated_outputs):
r_loss = torch.mean((1-dr)**2)
g_loss = torch.mean(dg**2)
loss += (r_loss + g_loss)
r_losses.append(r_loss.item())
g_losses.append(g_loss.item())
return loss, r_losses, g_losses
def generator_loss(disc_outputs):
loss = 0
gen_losses = []
for dg in disc_outputs:
l = torch.mean((1-dg)**2)
gen_losses.append(l)
loss += l
return loss, gen_losses

58
hifigan/utils.py Normal file
View File

@ -0,0 +1,58 @@
import glob
import os
import matplotlib
import torch
from torch.nn.utils import weight_norm
matplotlib.use("Agg")
import matplotlib.pylab as plt
def plot_spectrogram(spectrogram):
fig, ax = plt.subplots(figsize=(10, 2))
im = ax.imshow(spectrogram, aspect="auto", origin="lower",
interpolation='none')
plt.colorbar(im, ax=ax)
fig.canvas.draw()
plt.close()
return fig
def init_weights(m, mean=0.0, std=0.01):
classname = m.__class__.__name__
if classname.find("Conv") != -1:
m.weight.data.normal_(mean, std)
def apply_weight_norm(m):
classname = m.__class__.__name__
if classname.find("Conv") != -1:
weight_norm(m)
def get_padding(kernel_size, dilation=1):
return int((kernel_size*dilation - dilation)/2)
def load_checkpoint(filepath, device):
assert os.path.isfile(filepath)
print("Loading '{}'".format(filepath))
checkpoint_dict = torch.load(filepath, map_location=device)
print("Complete.")
return checkpoint_dict
def save_checkpoint(filepath, obj):
print("Saving checkpoint to {}".format(filepath))
torch.save(obj, filepath)
print("Complete.")
def scan_checkpoint(cp_dir, prefix):
pattern = os.path.join(cp_dir, prefix + '????????')
cp_list = glob.glob(pattern)
if len(cp_list) == 0:
return None
return sorted(cp_list)[-1]

16
toolbox/__init__.py
View File

@ -1,7 +1,8 @@
from toolbox.ui import UI
from encoder import inference as encoder
from synthesizer.inference import Synthesizer
from vocoder import inference as vocoder
from vocoder import inference as rnn_vocoder
from hifigan import inference as gan_vocoder
from pathlib import Path
from time import perf_counter as timer
from toolbox.utterance import Utterance
@ -13,6 +14,9 @@ import librosa
import re
from audioread.exceptions import NoBackendError
# 默认使用wavernn
vocoder = rnn_vocoder
# Use this directory structure for your datasets, or modify it to fit your needs
recognized_datasets = [
"LibriSpeech/dev-clean",
@ -353,10 +357,20 @@ class Toolbox:
self.ui.set_loading(0)
def init_vocoder(self):
global vocoder
model_fpath = self.ui.current_vocoder_fpath
# Case of Griffin-lim
if model_fpath is None:
return
# Sekect vocoder based on model name
if model_fpath.name[0] == "g":
vocoder = gan_vocoder
self.ui.log("vocoder is hifigan")
else:
vocoder = rnn_vocoder
self.ui.log("Loading the vocoder %s... " % model_fpath)
self.ui.set_loading(1)

BIN
vocoder/saved_models/pretrained/g_hifigan.pt Normal file
View File

Binary file not shown.