-
Notifications
You must be signed in to change notification settings - Fork 4
/
Train_RNN.py
367 lines (291 loc) · 13.9 KB
/
Train_RNN.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
import os
import argparse
import time
import shutil
import warnings
from logger import Logger
import random
import numpy as np
from sklearn.metrics import roc_auc_score
import torch
import torch.utils.data as data
import torch.nn as nn
import torch.optim as optim
from torch.autograd import Variable
import torch.backends.cudnn as cudnn
import config as c
from utils import read_feature, save_lr_and_losses, visualize_the_learning_rate, visualize_the_loss
from DB_reader import read_DB_structure
from Model import Model
from VAD_Dataset import VAD_Dataset, VAD_Compose, LSTMInputTrain, LSTMInputTest, ToTensorInput
from focalloss import focal_loss
os.environ['CUDA_VISIBLE_DEVICES'] = '0, 1, 2'
parser = argparse.ArgumentParser(description='PyTorch VAD(Voice Activity Detection)')
# Model Options
parser.add_argument('--log-dir', default='./checkpoints_', help='folder to output model checkpoints')
parser.add_argument('--resume', default=None, type=str, metavar='PATH',
help='path to latest checkpoint')
parser.add_argument('--workers', default=4, type=int, metavar='N',
help='number of data loading workers')
parser.add_argument('--start-epoch', default=1, type=int, metavar='SE',
help='manual epoch number')
parser.add_argument('--epochs', default=20, type=int, metavar='E',
help='number of epochs to train')
parser.add_argument('--RNN-model', default='BasicRNN', type=str, metavar='RM',
help='choose the RNN Model(BasicRNN or AttentionRNN)')
parser.add_argument('--attention-type', default='Combined', type=str, metavar='AT',
help='choose the attention type (TA / FA / DA1 / DA2)')
parser.add_argument('--hidden-size', default=64, type=int, metavar='HS',
help='number of LSTM hidden units')
parser.add_argument('--seq-len', default=50, type=int, metavar='SL',
help='number of sequential length')
parser.add_argument('--seed', default=2019, type=int, metavar='S',
help='random seed for initializing training')
parser.add_argument('--shuffle', action='store_true', default=True,
help='shuffle or not')
parser.add_argument('--num-layers', default=3, type=int, metavar='NL',
help='number of hidden layers')
parser.add_argument('--batch-size', default=128, type=int, metavar='BS',
help='input batch size for training')
parser.add_argument('--padding-time', default=1.0, type=float, metavar='PT',
help='padding time in train(valid) data')
parser.add_argument('--loss', default='FocalLoss', type=str, metavar='L',
help='choose the loss function. CrossEntropy or FocalLoss')
parser.add_argument('--valid-batch-size', default=1, type=int, metavar='VBS',
help='input batch size for validating(testing)')
parser.add_argument('--gamma', default=0.1, type=float, metavar='G',
help='hyper parameter for focal loss')
parser.add_argument('--lr', default=1e-1, type=float, metavar='LR',
help='starting learning rate')
parser.add_argument('--lr-decay', default=1e-4, type=float, metavar='LRD',
help='learning rate decay ratio')
parser.add_argument('--weight-decay', default=0.0, type=float, metavar='WD',
help='weight decay')
parser.add_argument('--optimizer', default='sgd', type=str, metavar='OPT',
help='the optimizer to use')
#Device
parser.add_argument('--no-cuda', default=False, action='store_true',
help='enables CUDA training')
parser.add_argument('--gpu-id', default='0', type=str,
help='id(s) for CUDA_VISIBLE_DEVICES')
parser.add_argument('--log-interval', default=22, metavar='LI',
help='how many batches to wait before logging training status')
args = parser.parse_args()
def save_checkpoint(state, is_best, filename='chekpoint.pth.tar'):
torch.save(state, filename)
if is_best:
shutil.copy(filename, 'model_best.pth.tar')
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self):
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
def create_optimizer(model, new_lr):
if args.optimizer == 'sgd':
optimizer = optim.SGD(model.parameters(), lr=new_lr, momentum=0.9, dampening=0, weight_decay=args.weight_decay)
elif args.optimizer == 'adam':
optimizer = optim.Adam(model.parameters(), lr=new_lr, weight_decay=args.wd)
elif args.optimizer == 'adagrad':
optimizer = optim.Adagrad(model.parameters(), lr=new_lr, lr_decay=args.lr_decay, weight_decay = args.wd)
return optimizer
def accuracy(output, target, topk=(1,)):
'''Computes the accuracy over the k top predictions for the specified values of k'''
with torch.no_grad():
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
res = []
for k in topk:
correct_k = correct[:k].view(-1).float().sum(0, keepdim=True)
res.append(correct_k.mul_(100.0 / batch_size))
return res
def get_learning_rate(optimizer):
lr = []
for param_group in optimizer.param_groups:
lr += [param_group['lr']]
return lr
def train(train_loader, model, criterion, optimizer, epoch):
batch_time = AverageMeter()
data_time = AverageMeter()
losses = AverageMeter()
n_correct, n_total = 0, 0
zero_count, one_count = 0, 0
model.train()
end = time.time()
for batch_idx, (data) in enumerate(train_loader):
inputs, targets = data
total_element = targets.shape[0] * targets.shape[1]
one_element = np.count_nonzero(targets)
zero_element = total_element - one_element
zero_count += zero_element
one_count += one_element
data_time.update(time.time() - end)
inputs = Variable(inputs)
targets = Variable(targets)
device_num = 'cuda:' + args.gpu_id
device = torch.device(device_num)
if args.cuda:
inputs, targets = inputs.to(device), targets.to(device)
linear_out, sigmoid_out = model(x=inputs)
pred = sigmoid_out >= 0.5
targets = targets.squeeze(-1)
n_correct += (pred.long() == targets.long()).sum().item()
n_total += args.seq_len * args.batch_size
train_acc = 100. * n_correct / n_total
loss = criterion(linear_out, targets.float())
losses.update(loss.item(), inputs.size(0))
optimizer.zero_grad()
loss.backward()
optimizer.step()
batch_time.update(time.time() - end)
end = time.time()
if batch_idx % args.log_interval == 0:
print(
'Train Epoch: {:3d} [{:8d}/{:8d} ({:3.0f}%)]\t'
'Time {batch_time.val:.3f} ({batch_time.avg:.3f})\t'
'Data {data_time.val:.3f} ({data_time.avg:.3f})\t'
'Loss {loss.val:.4f} ({loss.avg:.4f})\t'
'Acc {train_acc:.4f}'.format(
epoch, batch_idx * len(inputs), len(train_loader.dataset),
100. * batch_idx / len(train_loader), batch_time=batch_time,
data_time=data_time, loss=losses, train_acc=train_acc)
)
zero_proportion = (zero_count / (zero_count + one_count)) * 100
one_proportion = (one_count / (zero_count + one_count)) * 100
print('zero_element : {} zero_proportion : {}\n'
'one_element : {} one_proportion : {}\n'.format(zero_count, zero_proportion, one_count, one_proportion))
return losses.avg
def validate(val_loader, model, criterion, args):
batch_time = AverageMeter()
losses = AverageMeter()
mean_AUC = 0.
n_correct, n_total = 0., 0.
model.eval()
with torch.no_grad():
end = time.time()
for i, data in enumerate(val_loader):
inputs, targets = data
inputs = inputs.squeeze(0)
targets = targets.squeeze(0)
device_num = 'cuda:' + args.gpu_id
device = torch.device(device_num)
if args.cuda:
inputs, targets = inputs.to(device), targets.to(device)
linear_out, sigmoid_out = model(x=inputs)
linear_out[linear_out != linear_out] = 0
sigmoid_out[sigmoid_out != sigmoid_out] = 0
linear_out, sigmoid_out = linear_out.squeeze(0), sigmoid_out.squeeze(0)
loss = criterion(linear_out, targets.float())
np_targets = targets.data.cpu().numpy()
np_sigmoid_out = sigmoid_out.data.cpu().numpy()
np_sigmoid_out = np.nan_to_num(np_sigmoid_out)
temp_AUC = roc_auc_score(np_targets, np_sigmoid_out)
pred = sigmoid_out >= 0.5
n_correct += (pred.long() == targets.long()).sum().item()
n_total += len(targets)
val_acc = 100. * n_correct / n_total
losses.update(loss.item(), inputs.size(0))
mean_AUC += temp_AUC
batch_time.update(time.time() - end)
end = time.time()
if i % (args.log_interval * 20) == 0:
print('Validating the model: ({:8d}/{:8d})'.format(i, len(val_loader.dataset)))
mean_AUC /= (i + 1)
print(' * Validation => '
'Loss {loss.val:.4f} ({loss.avg:.4f})\t'
'AUC {mean_AUC:.3f}\t'
'Acc {val_acc:.5f}'.format(loss=losses, mean_AUC=mean_AUC * 100, val_acc=val_acc))
return losses.avg, mean_AUC
def main():
args.cuda = not args.no_cuda and torch.cuda.is_available()
device_num = 'cuda:' + args.gpu_id
device = torch.device(device_num)
if args.seed is not None:
random.seed(args.seed)
torch.manual_seed(args.seed)
np.random.seed(args.seed)
cudnn.deterministic = True
warnings.warn('You have chosen to seed training. '
'This will turn on the CUDNN deterministic setting, '
'which can slow down your training considerably! '
'You may see unexpected behavior when restarting '
'from checkpoints.')
if not os.path.exists(args.log_dir):
os.makedirs(args.log_dir)
LOG_DIR = args.log_dir + str(
args.seed) + '/Padding-{}/Atype-{}_Loss-{}_gamma-{}'.format(args.padding_time, args.attention_type, args.loss, args.gamma)
if not os.path.exists(LOG_DIR):
logger = Logger(LOG_DIR)
if args.cuda:
cudnn.benchmark = True
input_size = c.FILTER_BANK
model = Model(rnn_model=args.RNN_model, input_size=input_size, rnn_hidden_size=args.hidden_size,
num_layers=args.num_layers, dnn_hidden_size=c.P_DNN_HIDDEN_SIZE, seq_len=args.seq_len,
attention_type=args.attention_type)
train_feat_dir = os.path.join(c.MFB_DIR + '_' + str(float(args.padding_time)), 'train_folder')
valid_feat_dir = os.path.join(c.MFB_DIR + '_' + str(1.0), 'valid_folder')
train_DB = read_DB_structure(train_feat_dir, 'train')
valid_DB = read_DB_structure(valid_feat_dir, 'valid')
transform = VAD_Compose([
LSTMInputTrain(sequence_length=args.seq_len),
ToTensorInput()
])
transform_v = VAD_Compose([
LSTMInputTest(),
ToTensorInput()
])
file_loader = read_feature
train_dataset = VAD_Dataset(DB=train_DB, loader=file_loader, transform=transform)
valid_dataset = VAD_Dataset(DB=valid_DB, loader=file_loader, transform=transform_v)
print('\nParsed Options:\n{}\n'.format(vars(args)))
if args.cuda:
model.to(device)
start = args.start_epoch
end = start + args.epochs
if args.loss == 'CrossEntropy':
criterion = nn.BCEWithLogitsLoss()
elif args.loss == 'FocalLoss':
criterion = focal_loss(alpha=1.0, gamma=args.gamma)
optimizer = create_optimizer(model, args.lr)
scheduler = optim.lr_scheduler.ReduceLROnPlateau(optimizer, 'min', patience=1, min_lr=1e-5, verbose=True)
if args.resume:
if os.path.isfile(args.resume):
print('=> loading checkpoint {}'.format(args.resume))
checkpoint = torch.load(args.resume)
args.start_epoch = checkpoint['epoch']
checkpoint = torch.load(args.resume)
model.load_state_dict(checkpoint['state_dict'])
optimizer.load_state_dict(checkpoint['optimizer'])
else:
print('=> no checkpoint found at {}'.format(args.resume))
train_loader = torch.utils.data.DataLoader(dataset=train_dataset, batch_size=args.batch_size,
shuffle=args.shuffle, num_workers=args.workers, pin_memory=args.cuda)
valid_loader = torch.utils.data.DataLoader(dataset=valid_dataset, batch_size=args.valid_batch_size,
shuffle=False, num_workers=args.workers, pin_memory=args.cuda)
for epoch in range(start, end):
train_loss = train(train_loader, model, criterion, optimizer, epoch)
valid_loss, valid_AUC = validate(valid_loader, model, criterion, args)
scheduler.step(valid_loss, epoch)
current_LR = get_learning_rate(optimizer)[0]
print(' * Learning Rate : %0.4f' % current_LR)
save_lr_and_losses(LOG_DIR, epoch, current_LR, train_loss, valid_loss, valid_AUC)
torch.save({'epoch': epoch + 1, 'state_dict': model.state_dict(),
'optimizer': optimizer.state_dict()},
'{}/checkpoint {}.pth'.format(LOG_DIR, epoch))
min_loss, max_AUC = visualize_the_loss(LOG_DIR)
visualize_the_learning_rate(LOG_DIR)
model_total_params = sum(p.numel() for p in model.parameters())
print('The number of parameters = %d' % model_total_params)
if __name__ == '__main__':
main()