File size: 8,296 Bytes
b53f252 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 |
import os
import errno
import numpy as np
from copy import deepcopy
from miscc.config import cfg
from scipy.io.wavfile import write
from torch.nn import init
import torch
import torch.nn as nn
import torchvision.utils as vutils
from wavefile import WaveWriter, Format
import RT60
from multiprocessing import Pool
#############################
def KL_loss(mu, logvar):
# -0.5 * sum(1 + log(sigma^2) - mu^2 - sigma^2)
KLD_element = mu.pow(2).add_(logvar.exp()).mul_(-1).add_(1).add_(logvar)
KLD = torch.mean(KLD_element).mul_(-0.5)
return KLD
def compute_discriminator_loss(netD, real_RIRs, fake_RIRs,
real_labels, fake_labels,
conditions, gpus):
criterion = nn.BCELoss()
batch_size = real_RIRs.size(0)
cond = conditions.detach()
fake = fake_RIRs.detach()
real_features = nn.parallel.data_parallel(netD, (real_RIRs), gpus)
fake_features = nn.parallel.data_parallel(netD, (fake), gpus)
# real pairs
#print("util conditions ",cond.size())
inputs = (real_features, cond)
real_logits = nn.parallel.data_parallel(netD.get_cond_logits, inputs, gpus)
errD_real = criterion(real_logits, real_labels)
# wrong pairs
inputs = (real_features[:(batch_size-1)], cond[1:])
wrong_logits = \
nn.parallel.data_parallel(netD.get_cond_logits, inputs, gpus)
errD_wrong = criterion(wrong_logits, fake_labels[1:])
# fake pairs
inputs = (fake_features, cond)
fake_logits = nn.parallel.data_parallel(netD.get_cond_logits, inputs, gpus)
errD_fake = criterion(fake_logits, fake_labels)
if netD.get_uncond_logits is not None:
real_logits = \
nn.parallel.data_parallel(netD.get_uncond_logits,
(real_features), gpus)
fake_logits = \
nn.parallel.data_parallel(netD.get_uncond_logits,
(fake_features), gpus)
uncond_errD_real = criterion(real_logits, real_labels)
uncond_errD_fake = criterion(fake_logits, fake_labels)
#
errD = ((errD_real + uncond_errD_real) / 2. +
(errD_fake + errD_wrong + uncond_errD_fake) / 3.)
errD_real = (errD_real + uncond_errD_real) / 2.
errD_fake = (errD_fake + uncond_errD_fake) / 2.
else:
errD = errD_real + (errD_fake + errD_wrong) * 0.5
return errD, errD_real.data, errD_wrong.data, errD_fake.data
# return errD, errD_real.data[0], errD_wrong.data[0], errD_fake.data[0]
def compute_generator_loss(epoch,netD,real_RIRs, fake_RIRs, real_labels, conditions, gpus):
criterion = nn.BCELoss()
loss = nn.L1Loss() #nn.MSELoss()
loss1 = nn.MSELoss()
RT_error = 0
# print("num", real_RIRs.size(),real_RIRs.size()[0])
# input("kk")
cond = conditions.detach()
fake_features = nn.parallel.data_parallel(netD, (fake_RIRs), gpus)
# fake pairs
inputs = (fake_features, cond)
fake_logits = nn.parallel.data_parallel(netD.get_cond_logits, inputs, gpus)
MSE_error = loss(real_RIRs,fake_RIRs)
MSE_error1 = loss1(real_RIRs,fake_RIRs)
sample_size = real_RIRs.size()[0]
channel = 12
fs = 16000
rn = np.random.randint(sample_size-(channel*2))
real_wave = np.array(real_RIRs[rn:rn+channel].to("cpu").detach())
real_wave = real_wave.reshape(channel,4096)
fake_wave = np.array(fake_RIRs[rn:rn+channel].to("cpu").detach())
fake_wave = fake_wave.reshape(channel,4096)
pool = Pool(processes=12)
results =[]
for n in range(channel):
results.append(pool.apply_async(RT60.t60_parallel, args=(n,real_wave,fake_wave,fs,)))
T60_error =0
for result in results:
T60_error = T60_error + result.get()
RT_error = T60_error/channel
pool.close()
pool.join()
# T60_error =0
# for m in range(channel):
# real_wave_single = real_wave[:,(rn+m)]
# fake_wave_single = fake_wave[:,(rn+m)]
# Real_T60_val = RT60.t60_impulse(real_wave_single,fs)
# Fake_T60_val = RT60.t60_impulse(fake_wave_single,fs)
# T60_diff = abs(Real_T60_val-Fake_T60_val)
# T60_error = T60_error + T60_diff
# RT_error = T60_error/channel
# r = WaveWriter("real.wav", channels=portion, samplerate=fs)
# r.write(np.array(real_IR))
# f = WaveWriter("fake.wav", channels=portion, samplerate=fs)
# f.write(np.array(fake_IR))
# result = call_python_version("3.8", "RT60", "t60_error",
# ["real.wav","fake.wav"])
# # print("RT_error ",result)
# RT_error = float(result)
# print("RT_error ",RT_error)
# if(epoch<100):
# errD_fake = criterion(fake_logits, real_labels)# + 2* 4096 * MSE_error
# else:
# errD_fake = criterion(fake_logits, real_labels) + 2* 4096 * MSE_error
errD_fake = criterion(fake_logits, real_labels) + 5* 4096 * MSE_error1 + 40 * RT_error
if netD.get_uncond_logits is not None:
fake_logits = \
nn.parallel.data_parallel(netD.get_uncond_logits,
(fake_features), gpus)
uncond_errD_fake = criterion(fake_logits, real_labels)
errD_fake += uncond_errD_fake
return errD_fake, MSE_error,RT_error
#############################
def weights_init(m):
classname = m.__class__.__name__
if classname.find('Conv') != -1:
m.weight.data.normal_(0.0, 0.02)
elif classname.find('BatchNorm') != -1:
m.weight.data.normal_(1.0, 0.02)
m.bias.data.fill_(0)
elif classname.find('Linear') != -1:
m.weight.data.normal_(0.0, 0.02)
if m.bias is not None:
m.bias.data.fill_(0.0)
#############################
def save_RIR_results(data_RIR, fake, epoch, RIR_dir):
num = cfg.VIS_COUNT
fake = fake[0:num]
# data_RIR is changed to [0,1]
if data_RIR is not None:
data_RIR = data_RIR[0:num]
for i in range(num):
# #print("came 1")
real_RIR_path = RIR_dir+"/real_sample"+str(i)+".wav"
fake_RIR_path = RIR_dir+"/fake_sample"+str(i)+"_epoch_"+str(epoch)+".wav"
fs =16000
real_IR = np.array(data_RIR[i].to("cpu").detach())
fake_IR = np.array(fake[i].to("cpu").detach())
# #print("fake_IR ", fake_IR.size)
# #print("real_IR ", real_IR.size)
# #print("max real_IR ", max(real_IR[0]))
# #print("min real_IR ", min(real_IR[0]))
r = WaveWriter(real_RIR_path, channels=1, samplerate=fs)
r.write(np.array(real_IR))
f = WaveWriter(fake_RIR_path, channels=1, samplerate=fs)
f.write(np.array(fake_IR))
# write(real_RIR_path,fs,real_IR)
# write(fake_RIR_path,fs,fake_IR)
# write(real_RIR_path,fs,real_IR)
# write(fake_RIR_path,fs,fake_IR)
# vutils.save_image(
# data_RIR, '%s/real_samples.png' % RIR_dir,
# normalize=True)
# # fake.data is still [-1, 1]
# vutils.save_image(
# fake.data, '%s/fake_samples_epoch_%03d.png' %
# (RIR_dir, epoch), normalize=True)
else:
for i in range(num):
# #print("came 2")
fake_RIR_path = RIR_dir+"/small_fake_sample"+str(i)+"_epoch_"+str(epoch)+".wav"
fs =16000
fake_IR = np.array(fake[i].to("cpu").detach())
f = WaveWriter(fake_RIR_path, channels=1, samplerate=fs)
f.write(np.array(fake_IR))
# write(fake_RIR_path,fs,fake[i].astype(np.float32))
# vutils.save_image(
# fake.data, '%s/lr_fake_samples_epoch_%03d.png' %
# (RIR_dir, epoch), normalize=True)
def save_model(netG, netD, epoch, model_dir):
torch.save(
netG.state_dict(),
'%s/netG_epoch_%d.pth' % (model_dir, epoch))
torch.save(
netD.state_dict(),
'%s/netD_epoch_last.pth' % (model_dir))
#print('Save G/D models')
def mkdir_p(path):
try:
os.makedirs(path)
except OSError as exc: # Python >2.5
if exc.errno == errno.EEXIST and os.path.isdir(path):
pass
else:
raise
|