-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathtrain.py
132 lines (78 loc) · 3.93 KB
/
train.py
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
import pytorch_lightning as pl
import torch
import torch.nn.functional as F
import torchaudio.transforms as T
import torchaudio as ta
import os
from MelGAN import Discriminator_MelGAN
from SEANet import SEANet
from utils import *
from pesq import pesq
class RTBWETrain(pl.LightningModule):
def __init__(self, config):
super(RTBWETrain, self).__init__()
self.config = config
self.lr = config['optim']['learning_rate']
self.B1 = config['optim']['B1']
self.B2 = config['optim']['B2']
self.resampler = T.Resample(8000, 16000)
self.generator = SEANet(min_dim = 8, causality = True)
self.output_dir_path = config['train']['output_dir_path']
self.epoch_save_start = config['train']['epoch_save_start']
self.val_epoch = config['train']['val_epoch']
self.path_dir_bwe_pred = config['predict']['pred_output_path']
self.discriminator = Discriminator_MelGAN()
self.automatic_optimization = False
def forward(self,x):
x = self.resampler(x)
output = self.generator(x)
return output
def configure_optimizers(self):
optimizer_d = torch.optim.Adam(self.discriminator.parameters(), lr=self.lr, betas = (self.B1, self.B2))
optimizer_g = torch.optim.Adam(self.generator.parameters(), lr=self.lr, betas = (self.B1, self.B2))
return optimizer_d, optimizer_g #, [lrscheduler_d, lr_scheduler_g])
def training_step(self, batch, batch_idx):
optimizer_d, optimizer_g = self.optimizers()
wav_nb, wav_wb, _ = batch
wav_bwe = self.forward(wav_nb)
#optimize discriminator
self.toggle_optimizer(optimizer_d)
loss_d =self.discriminator.loss_D(wav_bwe, wav_wb)
optimizer_d.zero_grad()
self.manual_backward(loss_d)
optimizer_d.step()
self.untoggle_optimizer(optimizer_d)
#optimize generator
self.toggle_optimizer(optimizer_g)
loss_g = self.discriminator.loss_G(wav_bwe, wav_wb)
optimizer_g.zero_grad()
self.manual_backward(loss_g)
optimizer_g.step()
self.untoggle_optimizer(optimizer_g)
self.log("train_loss_d", loss_d, prog_bar = True, batch_size = self.config['dataset']['batch_size'])
self.log("train_loss_g", loss_g, prog_bar = True, batch_size = self.config['dataset']['batch_size'])
def validation_step(self, batch, batch_idx):
wav_nb, wav_wb, filename = batch
wav_bwe = self.forward(wav_nb)
loss_d = self.discriminator.loss_D(wav_bwe, wav_wb)
loss_g = self.discriminator.loss_G(wav_bwe, wav_wb)
wav_bwe_cpu = wav_bwe.squeeze(0).cpu()
val_dir_path = f"{self.output_dir_path}/epoch_current"
check_dir_exist(val_dir_path)
ta.save(os.path.join(val_dir_path, f"{filename[0]}.wav"), wav_bwe_cpu, 16000)
wav_wb = wav_wb.squeeze().cpu().numpy()
wav_bwe = wav_bwe.squeeze().cpu().numpy()
val_pesq_wb = pesq(fs = 16000, ref = wav_wb, deg = wav_bwe, mode = "wb")
val_pesq_nb = pesq(fs = 16000, ref = wav_wb, deg = wav_bwe, mode = "nb")
self.log_dict({"val_loss/val_loss_d": loss_d, "val_loss/val_loss_g": loss_g}, batch_size = 1, sync_dist=True)
self.log('val_pesq_wb', val_pesq_wb, batch_size = 1, sync_dist=True)
self.log('val_pesq_nb', val_pesq_nb, batch_size = 1, sync_dist=True)
def test_step(self, batch, batch_idx):
pass
def predict_step(self, batch, batch_idx):
wav_nb, _, filename = batch
wav_bwe = self.forward(wav_nb)
wav_bwe_cpu = wav_bwe.squeeze(0).cpu()
test_dir_path = self.path_dir_bwe_pred
check_dir_exist(test_dir_path)
ta.save(os.path.join(test_dir_path, f"{filename}.wav"), wav_bwe_cpu, 16000)