-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
676 lines (536 loc) · 24.4 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
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
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
# coding: utf-8
"""Trainining script for Tacotron speech synthesis model.
usage: train.py [options]
options:
--hparams_name=<name> Name of hyper params [default: vc].
--hparams=<parmas> Hyper parameters to be overrided [default: ].
--checkpoint-g=<name> Load generator from checkpoint if given.
--checkpoint-d=<name> Load discriminator from checkpoint if given.
--checkpoint-r=<name> Load reference model to compute spoofing rate.
--max_files=<N> Max num files to be collected. [default: -1]
--discriminator-warmup Warmup discriminator.
--w_d=<f> Adversarial loss weight [default: 1.0].
--mse_w=<f> Mean squared error (MSE) loss weight [default: 0.0].
--restart_epoch=<N> Restart epoch [default: -1].
--reset_optimizers Reset optimizers, otherwise restored from checkpoint.
--log-event-path=<name> Log event path.
--checkpoint-dir=<dir> Directory where to save model checkpoints [default: checkpoints].
--checkpoint-path=<name> Restore model from checkpoint path if given.
-h, --help Show this help message and exit
"""
import torch
import pickle
import numpy as np
from docopt import docopt
from torch import nn, optim
from torch.autograd import Variable
from torch.nn import functional as F
import torch.backends.cudnn as cudnn
from torch.utils import data as data_utils
from torch.autograd import grad as torch_grad
from sklearn.model_selection import train_test_split
import os
import sys
import math
import time
from tqdm import tqdm
from warnings import warn
from os.path import splitext, join, abspath, exists
# Make the tacotron files accessible
tacotron_lib_dir = join( expanduser("~"), "tacotron", "lib", "tacotron" )
sys.path.append( tacotron_lib_dir )
from tacotron_pytorch import Tacotron
from util import audio
from util.plot import plot_alignment
import tensorboard_logger
from tensorboard_logger import log_value
import gantts_models
from nnmnkwii import preprocessing as P
from nnmnkwii.datasets import FileSourceDataset, FileDataSource
from hparams import hparams
from text import text_to_sequence, symbols
from hparams_gan import hparams_debug_string
from hparams_gan import hparams as hparams_gan
hp = None # to be initailized later
gp_weight = 10
global_epoch = 0
global_step_gen = 0
checkpoint_interval = 10
use_cuda = torch.cuda.is_available()
DATA_ROOT = join(expanduser("~"), "tacotron", "training_ljspeech_gantts")
def gradient_penalty(model_d, real_data, generated_data, lengths):
batch_size = real_data.size()[0]
# Calculate interpolation
alpha = torch.rand(batch_size, 1, 1)
alpha = alpha.expand_as(real_data)
if use_cuda:
alpha = alpha.cuda()
interpolated = alpha * real_data.data + (1 - alpha) * generated_data.data
interpolated = Variable(interpolated, requires_grad=True)
if use_cuda:
interpolated = interpolated.cuda()
# Calculate probability of interpolated examples
prob_interpolated = model_d(interpolated, lengths=lengths)
# Calculate gradients of probabilities with respect to examples
gradients = torch_grad(outputs=prob_interpolated, inputs=interpolated,
grad_outputs=torch.ones(prob_interpolated.size()).cuda() if use_cuda else torch.ones(
prob_interpolated.size()),
create_graph=True, retain_graph=True)[0]
# Flatten Gradiengts to easily take norm per example in batch
gradients = gradients.contiguous()
gradients = gradients.view(batch_size, -1)
# Derivatives of the gradient close to 0 can cause problems because of
# the square root, so manually calculate norm and add epsilon
gradients_norm = torch.sqrt(torch.sum(gradients ** 2, dim=1) + 1e-12)
# Return gradient penalty
return gp_weight * ((gradients_norm - 1) ** 2).mean()
# Pad all inputs with zeros to make them the same length
def _pad_2d(x, max_len):
print(np.array(x).shape)
print(max_len)
x = np.pad(x, [(0, max_len - len(x)), (0, 0)],
mode="constant", constant_values=0)
return x
class TextDataSource(FileDataSource):
def __init__(self):
self._cleaner_names = [x.strip() fortqdm x in hparams.cleaners.split(',')]
def collect_files(self):
meta = join(DATA_ROOT, "train.txt")
with open(meta, "rb") as f:
lines = f.readlines()
lines = list(map(lambda l: l.decode("utf-8").split("|")[-1], lines))
return lines
def collect_features(self, text):
return np.asarray(text_to_sequence(text, self._cleaner_names),
dtype=np.int32)
class _NPYDataSource(FileDataSource):
def __init__(self, col):
self.col = col
def collect_files(self):
meta = join(DATA_ROOT, "train.txt")
with open(meta, "rb") as f:
lines = f.readlines()
lines = list(map(lambda l: l.decode("utf-8").split("|")[self.col], lines))
paths = list(map(lambda f: join(DATA_ROOT, f), lines))
return paths
def collect_features(self, path):
return np.load(path)
class MelSpecDataSource(_NPYDataSource):
def __init__(self):
super(MelSpecDataSource, self).__init__(1)
class LinearSpecDataSource(_NPYDataSource):
def __init__(self):
super(LinearSpecDataSource, self).__init__(0)
class PyTorchDataset(object):
def __init__(self, X, Mel, Y):
self.X = X
self.Mel = Mel
self.Y = Y
def __getitem__(self, idx):
return self.X[idx], self.Mel[idx], self.Y[idx]
def __len__(self):
return len(self.X)
def collate_fn(batch):
"""Create batch"""
r = hparams.outputs_per_step
input_lengths = [len(x[0]) for x in batch]
max_input_len = np.max(input_lengths)
# Add single zeros frame at least, so plus 1
max_target_len = np.max([len(x[1]) for x in batch]) + 1
if max_target_len % r != 0:
max_target_len += r - max_target_len % r
assert max_target_len % r == 0
a = np.array([_pad(x[0], max_input_len) for x in batch], dtype=np.int)
x_batch = torch.LongTensor(a)
input_lengths = torch.LongTensor(input_lengths)
b = np.array([_pad_2d(x[1], max_target_len) for x in batch],
dtype=np.float32)
mel_batch = torch.FloatTensor(b)
c = np.array([_pad_2d(x[2], max_target_len) for x in batch],
dtype=np.float32)
y_batch = torch.FloatTensor(c)
return x_batch, input_lengths, mel_batch, y_batch
def save_alignment(path, attn):
plot_alignment(attn.T, path, info="tacotron, step={}".format(global_step))
def save_spectrogram(path, linear_output):
spectrogram = audio._denormalize(linear_output)
plt.figure(figsize=(16, 10))
plt.imshow(spectrogram.T, aspect="auto", origin="lower")
plt.colorbar()
plt.tight_layout()
plt.savefig(path, format="png")
plt.close()
def save_states(global_step, linear_outputs, attn, y,
input_lengths, checkpoint_dir=None):
print("Save intermediate states at step {}".format(global_step))
idx = min(1, len(input_lengths) - 1)
input_length = input_lengths[idx]
# Alignment
path = join(checkpoint_dir, "step{}_alignment.png".format(
global_step))
alignment = attn[idx].cpu().data.numpy()
save_alignment(path, alignment)
# Predicted spectrogram
path = join(checkpoint_dir, "step{}_predicted_spectrogram.png".format(
global_step))
linear_output = linear_outputs[idx].cpu().data.numpy()
save_spectrogram(path, linear_output)
# Predicted audio signal
signal = audio.inv_spectrogram(linear_output.T)
path = join(checkpoint_dir, "step{}_predicted.wav".format(
global_step))
audio.save_wav(signal, path)
# Target spectrogram
path = join(checkpoint_dir, "step{}_target_spectrogram.png".format(
global_step))
linear_output = y[idx].cpu().data.numpy()
save_spectrogram(path, linear_output)
def save_checkpoint(model, optimizer, epoch, step, checkpoint_dir, name):
checkpoint_path = join(
checkpoint_dir, "checkpoint_epoch{}_{}.pth".format(
epoch, name))
torch.save({
"state_dict": model.state_dict(),
"optimizer": optimizer.state_dict(),
"global_step": step,
"global_epoch": epoch,
}, checkpoint_path)
print("Saved checkpoint:", checkpoint_path)
def update_discriminator(model_d, optimizer_d, x, y_static, y_hat, lengths,
mask, phase, eps=1e-20):
y_static_adv, y_hat_static_adv = y_static, y_hat
if hp.discriminator_linguistic_condition:
y_static_adv = torch.cat((x, y_static_adv), -1)
y_hat_static_adv = torch.cat((x, y_hat_static_adv), -1)
# Real
D_real = model_d(y_static_adv, lengths=lengths)
real_correct_count = (D_real > 0.5).float().item()
# Fake
D_fake = model_d(y_hat_static_adv, lengths=lengths)
fake_correct_count = (D_fake < 0.5).float().item()
# Wasserstein Loss
# Do we take an average
loss_real_d = - (D_real + eps)
loss_fake_d = - (D_fake + eps)
loss_d = loss_real_d - loss_fake_d
loss_d = loss_d + gradient_penalty(model_d, y_static_adv, y_hat_static_adv, lengths)
loss_d.backward(retain_graph=True)
torch.nn.utils.clip_grad_norm(model_d.parameters(), 1.0)
optimizer_d.step()
return loss_d.item(), loss_fake_d.item(), loss_real_d.item(), real_correct_count, fake_correct_count
def update_generator(model_g, model_d, optimizer_g,
x, y, mel, mel_hat, mel_static, linear_output,
adv_w, lengths, mask, phase,
mse_w=None, mge_w=None, eps=1e-20):
criterion = nn.L1Loss(size_average=False)
# Adversarial loss
if adv_w > 0:
# Select streams
mel_hat_static_adv = mel_hat
if hp.discriminator_linguistic_condition:
mel_hat_static_adv = torch.cat((x, mel_hat_static_adv), -1)
model_local_output = model_d( mel_hat_static_adv, lengths=lengths )
# No log in Wasserstein GAN
loss_adv = model_local_output + eps
else:
loss_adv = Variable(mel.data.new(1).zero_())
mel_loss = criterion( mel_hat, mel )
n_priority_freq = int(3000 / (fs * 0.5) * linear_dim)
linear_loss = 0.5 * criterion(y, linear_output) + 0.5 * criterion(y[:, :, :n_priority_freq], linear_output[:, :, :n_priority_freq])
# L-1 + ADV loss
gen_loss = linear_loss + mel_loss
loss_g = gen_loss + (adv_w * loss_adv)
loss_g.backward()
torch.nn.utils.clip_grad_norm(model_g.parameters(), 1.0)
optimizer_g.step()
return gen_loss.item(), loss_adv.item(), loss_g.item()
def _learning_rate_decay(init_lr, global_step):
warmup_steps = 4000.0
step = global_step + 1.
lr = init_lr * warmup_steps**0.5 * np.minimum(
step * warmup_steps**-1.5, step**-0.5)
return lr
def exp_lr_scheduler(optimizer, epoch, nepoch, init_lr=0.0001, lr_decay_epoch=100):
"""Decay learning rate by a factor of 0.1 every lr_decay_epoch epochs."""
lr = init_lr * (0.1**(epoch // lr_decay_epoch))
if epoch % lr_decay_epoch == 0:
print('LR is set to {} at epoch {}'.format(lr, epoch))
for param_group in optimizer.param_groups:
param_group['lr'] = lr
return optimizer
def apply_generator(model_g, x, R, mel, lengths):
if model_g.include_parameter_generation():
# Case: models include parameter generation in itself
# Mulistream features cannot be used in this case
mel_hat, mel_hat_static = model_g(x, mel, R, lengths=lengths)
else:
# Case: generic models (can be sequence model)
assert hp.has_dynamic_features is not None
mel_hat, linear_op, attn = model_g( x, mel, lengths=lengths )
# Handle dimention mismatch
# This happens when we use pad_packed_sequence
if mel_hat.size(1) != x.size(1):
mel_hat = F.pad(mel_hat.unsqueeze(
0), (0, 0, x.size(1) - mel_hat.size(-2), 0)).squeeze(0)
return mel_hat, linear_op, attn
def train_loop( models, optimizers, dataset_loaders, w_d=0.0, mse_w=0.0, mge_w=1.0, update_d=True, update_g=True ):
# Intial learning rate
init_lr = 0.002
model_g, model_d = models
optimizer_g, optimizer_d = optimizers
if use_cuda:
model_g, model_d = model_g.cuda(), model_d.cuda()
Y_data_mean = dataset_loaders["train"].dataset.Y_data_mean
Y_data_std = dataset_loaders["train"].dataset.Y_data_std
Y_data_mean = torch.from_numpy(Y_data_mean)
Y_data_std = torch.from_numpy(Y_data_std)
if use_cuda:
Y_data_mean = Y_data_mean.cuda()
Y_data_std = Y_data_std.cuda()
E_loss_mge = 1
E_loss_adv = 1
has_dynamic = np.any(hp.has_dynamic_features)
linear_outputs = []
attn = []
global global_epoch, global_step_gen
print("Running for "+str(hp.nepoch)+" epochs")
for global_epoch in range(global_epoch + 1, hp.nepoch + 1):
print("Epoch : "+str(global_epoch))
if hp.lr_decay_schedule and update_d:
optimizer_d = exp_lr_scheduler(optimizer_d, global_epoch - 1, hp.nepoch,
init_lr=hp.optimizer_d_params["lr"],
lr_decay_epoch=hp.lr_decay_epoch)
running_loss = {"generator": 0.0, "mse": 0.0,
"loss_real_d": 0.0,
"loss_fake_d": 0.0,
"loss_adv": 0.0,
"discriminator": 0.0}
model_g.train()
running_metrics = {}
real_correct_count, fake_correct_count = 0, 0
regard_fake_as_natural = 0
N = len(dataset_loaders)
total_num_frames = 0
# update_g_local implements More Disc iters per Gen iter.
if(update_g and update_d):
update_g_local = False
else:
update_g_local = True
for ind, (x, input_lengths, mel, y) in tqdm(enumerate(dataset_loaders)):
if ( update_g_local==False and ind!=0 and ind%5 ==0):
# We will be updating the gebnerator in this iteration. So update the learning rate according to the current number of generator step.
update_g_local=True
current_lr = _learning_rate_decay(init_lr, global_step_gen)
for param_group in optimizer_g.param_groups:
param_group['lr'] = current_lr
# Sort by lengths. This is needed for pytorch's PackedSequence
sorted_lengths, indices = torch.sort(
lengths.view(-1), dim=0, descending=True)
sorted_lengths = sorted_lengths.long()
cpu_sorted_lengths = list(sorted_lengths)
max_len = sorted_lengths[0]
# Get sorted batch
x, mel, y = x[indices], mel[indices], y[indices]
# Generator noise
if hp.generator_add_noise:
z = torch.rand(x.size(0), max_len, hp.generator_noise_dim)
else:
z = None
# Construct MLPG paramgen matrix for every batch
if has_dynamic:
R = unit_variance_mlpg_matrix(hp.windows, max_len)
R = torch.from_numpy(R)
R = R.cuda() if use_cuda else R
else:
R = None
if use_cuda:
x, y, mel = x.cuda(), y.cuda(), mel.cuda()
sorted_lengths = sorted_lengths.cuda()
z = z.cuda() if z is not None else None
# Pack into variables
x, y, mel = Variable(x), Variable(y), Variable(mel)
z = Variable(z) if z is not None else None
sorted_lengths = Variable(sorted_lengths)
# Static features
mel_static = get_static_features(
mel, len(hp.windows), hp.stream_sizes, hp.has_dynamic_features)
# Num frames in batch
total_num_frames += sorted_lengths.float().sum().item()
# Mask -- NEEDED?
(sorted_lengths).unsqueeze(-1)
mask = None
# Reset optimizers state
optimizer_g.zero_grad()
optimizer_d.zero_grad()
# Apply model (generator)
generator_input = torch.cat((x, z), -1) if z is not None else x
mel_hat, linear_outputs, attn = apply_generator( model_g, generator_input, R, mel, cpu_sorted_lengths )
### Update discriminator ###
# Natural: 1, Generated: 0
if update_d:
loss_d, loss_fake_d, loss_real_d, _real_correct_count,\
_fake_correct_count = update_discriminator(
model_d, optimizer_d, x, mel_static, mel_hat
cpu_sorted_lengths, mask, phase)
running_loss["discriminator"] += loss_d
running_loss["loss_fake_d"] += loss_fake_d
running_loss["loss_real_d"] += loss_real_d
real_correct_count += _real_correct_count
fake_correct_count += _fake_correct_count
### Update generator ###
if (update_g==True and update_d==False) or (update_g==True and update_d==True and update_g_local==True):
if (E_loss_adv == 0):
adv_w = w_d * 1e+3
else:
adv_w = w_d * float(np.clip(E_loss_mge / E_loss_adv, 0, 1e+3))
loss_tacotron, loss_adv, loss_g = update_generator(
model_g, model_d, optimizer_g, x, y, mel, mel_hat,
mel_static, mel_hat_static,linear_output,
adv_w, cpu_sorted_lengths, mask, phase,
mse_w=mse_w, mge_w=mge_w)
running_loss["generator"] += loss_g
running_loss["loss_adv"] += loss_adv
running_loss["tacotron"] += loss_tacotron
global_step_gen += 1
update_g_local = False
save_states( global_step, linear_outputs, attn, y, sorted_lengths, checkpoint_dir )
# Log loss
for ty, enabled in [("tacotron", update_g),
("discriminator", update_d),
("loss_real_d", update_d),
("loss_fake_d", update_d),
("loss_adv", update_g and update_d),
("generator", update_g)]:
if enabled:
ave_loss = running_loss[ty] / N
log_value(
"{} {} loss".format(phase, ty), ave_loss, global_epoch)
# Log discriminator classification accuracy
if update_d:
log_value("Real {} acc".format(phase),
real_correct_count / total_num_frames, global_epoch)
log_value("Fake {} acc".format(phase),
fake_correct_count / total_num_frames, global_epoch)
# Save checkpoints
if global_epoch % checkpoint_interval == 0:
for model, optimizer, enabled, name in [
(model_g, optimizer_g, update_g, "Generator"),
(model_d, optimizer_d, update_d, "Discriminator")]:
if enabled:
save_checkpoint( model, optimizer, global_epoch, global_step, checkpoint_dir, name )
return 0
def load_checkpoint(model, optimizer, checkpoint_path):
global global_epoch
print("Load checkpoint from: {}".format(checkpoint_path))
checkpoint = torch.load(checkpoint_path)
model.load_state_dict(checkpoint["state_dict"])
if optimizer is not None:
optimizer.load_state_dict(checkpoint["optimizer"])
global_epoch = checkpoint["global_epoch"]
global_epoch = checkpoint["global_step"]
if __name__ == "__main__":
since = time.time()
args = docopt(__doc__)
print("Command line args:\n", args)
hp = getattr(hparams_gan, args["--hparams_name"])
# Override hyper parameters
hp.parse(args["--hparams_gan"])
checkpoint_dir = args["--checkpoint-dir"]
checkpoint_path_d = args["--checkpoint-d"]
checkpoint_path_g = args["--checkpoint-g"]
checkpoint_path_r = args["--checkpoint-r"]
max_files = int(args["--max_files"])
w_d = float(args["--w_d"])
mse_w = float(args["--mse_w"])
mge_w = float(args["--mge_w"])
discriminator_warmup = args["--discriminator-warmup"]
restart_epoch = int(args["--restart_epoch"])
reset_optimizers = args["--reset_optimizers"]
log_event_path = args["--log-event-path"]
# Flags to update discriminator/generator or not
update_d = w_d > 0
update_g = False if discriminator_warmup else True
if not exists(checkpoint_dir):
os.makedirs(checkpoint_dir)
X = []
Y = []
# utt_lengths = []
X = FileSourceDataset(TextDataSource())
Mel = FileSourceDataset(MelSpecDataSource())
Y = FileSourceDataset(LinearSpecDataSource())
print("Size of dataset for {}: {}".format(phase, len(X)))
ty = "acoustic" if hp == hparams_gan.tts_acoustic else "duration"
X_data_min, X_data_max = P.minmax( X )
Mel_data_mean, Mel_data_var = P.meanvar( Mel )
Mel_data_std = np.sqrt( Mel_data_var )
np.save(join(data_dir, "X_{}_data_min".format(ty)), X_data_min)
np.save(join(data_dir, "X_{}_data_max".format(ty)), X_data_max)
np.save(join(data_dir, "Mel_{}_data_mean".format(ty)), Mel_data_mean)
np.save(join(data_dir, "Mel_{}_data_var".format(ty)), Mel_data_var)
if hp.discriminator_params["in_dim"] is None:
sizes = get_static_stream_sizes(
hp.stream_sizes, hp.has_dynamic_features, len(hp.windows))
D = int(np.array(sizes[hp.adversarial_streams]).sum())
if hp.adversarial_streams[0]:
D -= hp.mask_nth_mgc_for_adv_loss
if hp.discriminator_linguistic_condition:
D = D + X_data_min.shape[-1]
hp.discriminator_params["in_dim"] = D
dataset = PyTorchDataset(X, Mel, Y)
data_loader = data_utils.DataLoader(
dataset, batch_size=hparams.batch_size,
num_workers=hparams.num_workers, shuffle=True,
collate_fn=collate_fn, pin_memory=hparams.pin_memory)
# Models
model_g = Tacotron(n_vocab = len(symbols),
embedding_dim = 256,
mel_dim = hparams.num_mels,
linear_dim = hparams.num_freq,
r = hparams.outputs_per_step,
padding_idx = hparams.padding_idx,
use_memory_mask = hparams.use_memory_mask,
)
model_d = getattr(gantts_models, hp.discriminator)(**hp.discriminator_params)
print("Generator:", model_g)
print("Discriminator:", model_d)
if use_cuda:
model_g, model_d = model_g.cuda(), model_d.cuda()
# Optimizers
optimizer_g = optim.Adam( model.parameters(),
lr = hparams.initial_learning_rate, betas = ( hparams.adam_beta1, hparams.adam_beta2 ),
weight_decay = hparams.weight_decay )
optimizer_d = getattr(optim, hp.optimizer_d)(model_d.parameters(),
**hp.optimizer_d_params)
# Load checkpoint
if checkpoint_path_d:
if reset_optimizers:
load_checkpoint(model_d, None, checkpoint_path_d)
else:
load_checkpoint(model_d, optimizer_d, checkpoint_path_d)
if checkpoint_path_g:
if reset_optimizers:
load_checkpoint(model_g, None, checkpoint_path_g)
else:
load_checkpoint(model_g, optimizer_g, checkpoint_path_g)
# Restart iteration at restart_epoch
if restart_epoch >= 0:
global_epoch = restart_epoch
# Setup tensorboard logger
if log_event_path is None:
log_event_path = "log/run-test" + str(np.random.randint(100000))
print("Loss event path: {}".format(log_event_path))
tensorboard_logger.configure(log_event_path)
# Train
print("Start training from epoch {}".format(global_epoch))
train_loop((model_g, model_d), (optimizer_g, optimizer_d),
dataset_loaders, w_d=w_d, update_d=update_d, update_g=update_g,
mse_w=mse_w, mge_w=mge_w)
# Save models -- Training Done. Save whatever progress was made since the last checkpoint was saved.
for model, optimizer, enabled, name in [
(model_g, optimizer_g, update_g, "Generator"),
(model_d, optimizer_d, update_d, "Discriminator")]:
if enabled:
save_checkpoint(
model, optimizer, global_epoch, global_step, checkpoint_dir, name)
print("Finished!")
sys.exit(0)