/
xvector_DPlda_pytorch.py
executable file
·195 lines (147 loc) · 8.3 KB
/
xvector_DPlda_pytorch.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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Mon Apr 6 12:30:45 2020
@author: shreyasr
"""
# %% imports and definitions
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import numpy as np
import os
import sys
import random
import pickle
import subprocess
from utils.NpldaConf import NpldaConf
from pdb import set_trace as bp
from utils.sv_trials_loaders import combine_trials_and_get_loader, get_trials_loaders_dict, load_xvec_trials_from_numbatch, load_xvec_trials_from_idbatch
from datetime import datetime
import logging
from utils.models import DPlda
def train(nc, model, device, train_loader, mega_xvec_dict, num_to_id_dict, optimizer, epoch, valid_loaders=None):
model.train()
losses = []
for batch_idx, (data1, data2, target) in enumerate(train_loader):
optimizer.zero_grad()
data1, data2, target = data1.to(device), data2.to(device), target.to(device)
data1_xvec, data2_xvec = load_xvec_trials_from_numbatch(mega_xvec_dict, num_to_id_dict, data1, data2, device)
output = model(data1_xvec, data2_xvec)
loss = model.loss(output, target)
losses.append(loss.item())
loss.backward()
optimizer.step()
if batch_idx % nc.log_interval == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\t {}: {:.6f}'.format(
epoch, batch_idx * len(data1), len(train_loader.dataset),
100. * batch_idx / len(train_loader), nc.loss, sum(losses)/len(losses)))
logging.info('Train Epoch: {} [{}/{} ({:.0f}%)]\t {}: {:.6f}'.format(
epoch, batch_idx * len(data1), len(train_loader.dataset),
100. * batch_idx / len(train_loader), nc.loss, sum(losses)/len(losses)))
losses = []
def validate(nc, model, device, mega_xvec_dict, num_to_id_dict, data_loader, update_thresholds=False):
model.eval()
with torch.no_grad():
targets, scores = torch.tensor([]).to(device), torch.tensor([]).to(device)
for data1, data2, target in data_loader:
data1, data2, target = data1.to(device), data2.to(device), target.to(device)
data1_xvec, data2_xvec = load_xvec_trials_from_numbatch(mega_xvec_dict, num_to_id_dict, data1, data2,
device)
targets = torch.cat((targets, target))
scores_batch = model.forward(data1_xvec, data2_xvec)
scores = torch.cat((scores, scores_batch))
soft_cdet_loss = model.softcdet(scores, targets)
cdet_mdl = model.cdet(scores, targets)
minc, minc_threshold = model.minc(scores, targets, update_thresholds)
logging.info('\n\nTest set: C_det (mdl): {:.4f}\n'.format(cdet_mdl))
logging.info('Test set: soft C_det (mdl): {:.4f}\n'.format(soft_cdet_loss))
logging.info('Test set: C_min: {:.4f}\n'.format(minc))
for beta in nc.beta:
logging.info('Test set: argmin threshold [{}]: {:.4f}\n'.format(beta, minc_threshold[beta]))
print('\n\nTest set: C_det (mdl): {:.4f}\n'.format(cdet_mdl))
print('Test set: soft C_det (mdl): {:.4f}\n'.format(soft_cdet_loss))
print('Test set: C_min: {:.4f}\n'.format(minc))
for beta in nc.beta:
print('Test set: argmin threshold [{}]: {:.4f}\n'.format(beta, minc_threshold[beta]))
return minc, minc_threshold
# %% main_kaldiplda
def main_kaldiplda():
timestamp = int(datetime.timestamp(datetime.now()))
print(timestamp)
logging.basicConfig(filename='logs/kaldiplda_{}.log'.format(timestamp),
filemode='a',
format='%(levelname)s: %(message)s',
datefmt='%H:%M:%S',
level=logging.DEBUG)
# %% Configure Training
configfile = 'conf/voices_config_dplda.cfg'
nc = NpldaConf(configfile)
torch.manual_seed(nc.seed)
np.random.seed(nc.seed)
random.seed(nc.seed)
logging.info(" Running file {}\n\nStarted at {}.\n".format(sys.argv[0], datetime.now()))
if not torch.cuda.is_available():
nc.device='cpu'
device = torch.device(nc.device)
print("Running on {}...".format(nc.device))
logging.info("Running on {} ...\n".format(nc.device))
logging.info("\nConfiguration:\n\n{}\n\n".format(''.join(open(configfile,'r').readlines())))
# %%Load the generated training data trials and make loaders here
mega_xvec_dict = pickle.load(open(nc.mega_xvector_pkl, 'rb'))
num_to_id_dict = {i: j for i, j in enumerate(list(mega_xvec_dict))}
id_to_num_dict = {v: k for k, v in num_to_id_dict.items()}
train_loader = combine_trials_and_get_loader(nc.training_data_trials_list, id_to_num_dict, subsample_factors=nc.train_subsample_factors ,batch_size=nc.batch_size)
# train_loader_sampled = combine_trials_and_get_loader(nc.training_data_trials_list, id_to_num_dict, batch_size=nc.batch_size, subset=0.05)
valid_loaders_dict = get_trials_loaders_dict(nc.validation_trials_list, id_to_num_dict, subsample_factors=nc.valid_subsample_factors, batch_size=5*nc.batch_size)
# %% Initialize model and stuff
model = DPlda(nc).to(device)
## To load a Kaldi trained PLDA model, Specify the paths of 'mean.vec', 'transform.mat' and 'plda' generated from stage 8 of https://github.com/kaldi-asr/kaldi/blob/master/egs/sre16/v2/run.sh
if nc.initialization == 'kaldi':
model.LoadParamsFromKaldi(nc.meanvec, nc.transformmat)
## Uncomment to initialize with a pickled pretrained model
# model = pickle.load(open('/home/data2/SRE2019/shreyasr/X/models/kaldi_pldaNet_sre0410_swbd_16_16.swbdsremx6epoch.1571651491.pt','rb'))
params_dict = dict(model.named_parameters())
updatable_params = []
for param in params_dict.keys():
if 'centering_and_LDA' in param:
params_dict[param].requires_grad = False
else:
updatable_params.append(params_dict[param])
optimizer = optim.Adam(updatable_params, lr=nc.lr, weight_decay=1e-5)
# optimizer = optim.Adam(model.parameters(), lr=nc.lr, weight_decay=1e-5)
print("Initializing the thresholds... Whatever numbers that get printed here are junk.\n")
valloss, minC_threshold = validate(nc, model, device, mega_xvec_dict, num_to_id_dict, valid_loaders_dict[nc.heldout_set_for_th_init], update_thresholds=True)
# %% Train and Validate model
print("\n\nEpoch 0: After Initialization\n")
all_losses = []
for val_set, valid_loader in valid_loaders_dict.items():
print("Validating {}".format(val_set))
logging.info("Validating {}".format(val_set))
valloss, minC_threshold = validate(nc, model, device, mega_xvec_dict, num_to_id_dict, valid_loader)
if val_set==nc.heldout_set_for_lr_decay:
all_losses.append(valloss)
for epoch in range(1, nc.n_epochs + 1):
train(nc, model, device, train_loader , mega_xvec_dict, num_to_id_dict, optimizer, epoch)
for val_set, valid_loader in valid_loaders_dict.items():
print("Validating {}".format(val_set))
logging.info("Validating {}".format(val_set))
valloss, minC_threshold = validate(nc, model, device, mega_xvec_dict, num_to_id_dict, valid_loader)
if val_set==nc.heldout_set_for_lr_decay:
all_losses.append(valloss)
model.SaveModel("models/NPLDA_{}_{}.pt".format(epoch, timestamp))
for trial_file in nc.test_trials_list:
print("Generating scores for Epoch {} with trial file {}".format(epoch, trial_file))
nc.generate_scorefile("scores/kaldipldanet_epoch{}_{}_{}.txt".format(epoch, os.path.splitext(os.path.basename(trial_file))[0], timestamp), trial_file, mega_xvec_dict, model, device, 5*nc.batch_size)
try:
if (all_losses[-1] > all_losses[-2]) and (all_losses[-2] > all_losses[-3]):
nc.lr = nc.lr / 2
print("REDUCING LEARNING RATE to {} since loss trend looks like {}".format(nc.lr, all_losses[-3:]))
logging.info("REDUCING LEARNING RATE to {} since loss trend looks like {}".format(nc.lr, all_losses[-3:]))
optimizer = optim.Adam(model.parameters(), lr=nc.lr, weight_decay=1e-5)
except:
pass
# %% __main__
if __name__ == '__main__':
main_kaldiplda()