xvector.py 78.2 KB
Newer Older
Anthony Larcher's avatar
Anthony Larcher committed
1
# coding: utf-8 -*-
Anthony Larcher's avatar
Anthony Larcher committed
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
#
# This file is part of SIDEKIT.
#
# SIDEKIT is a python package for speaker verification.
# Home page: http://www-lium.univ-lemans.fr/sidekit/
#
# SIDEKIT is a python package for speaker verification.
# Home page: http://www-lium.univ-lemans.fr/sidekit/
#
# SIDEKIT is free software: you can redistribute it and/or modify
# it under the terms of the GNU LLesser General Public License as
# published by the Free Software Foundation, either version 3 of the License,
# or (at your option) any later version.
#
# SIDEKIT is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public License
# along with SIDEKIT.  If not, see <http://www.gnu.org/licenses/>.

"""
Anthony Larcher's avatar
v1.3.7    
Anthony Larcher committed
25
Copyright 2014-2021 Anthony Larcher, Yevhenii Prokopalo
Anthony Larcher's avatar
Anthony Larcher committed
26
"""
Anthony Larcher's avatar
Anthony Larcher committed
27

Anthony Larcher's avatar
Anthony Larcher committed
28
29
30

import pdb
import traceback
Anthony Larcher's avatar
Anthony Larcher committed
31
import logging
Anthony Larcher's avatar
debug    
Anthony Larcher committed
32
import math
Anthony Larcher's avatar
Anthony Larcher committed
33
import os
Anthony Larcher's avatar
Anthony Larcher committed
34
import numpy
Anthony Larcher's avatar
Anthony Larcher committed
35
import pandas
Anthony Larcher's avatar
minor    
Anthony Larcher committed
36
import pickle
Anthony Larcher's avatar
Anthony Larcher committed
37
import shutil
Anthony Larcher's avatar
Anthony Larcher committed
38
import sys
39
import time
Anthony Larcher's avatar
Anthony Larcher committed
40
import torch
Anthony Larcher's avatar
debug    
Anthony Larcher committed
41
import torchaudio
Anthony Larcher's avatar
Anthony Larcher committed
42
import tqdm
Anthony Larcher's avatar
Anthony Larcher committed
43
44
import yaml

Anthony Larcher's avatar
Anthony Larcher committed
45
from torchvision import transforms
Anthony Larcher's avatar
Anthony Larcher committed
46
from collections import OrderedDict
Anthony Larcher's avatar
Anthony Larcher committed
47
from .xsets import SideSet
Anthony Larcher's avatar
debug    
Anthony Larcher committed
48
from .xsets import FileSet
Anthony Larcher's avatar
Anthony Larcher committed
49
from .xsets import IdMapSet
Anthony Larcher's avatar
Anthony Larcher committed
50
from .xsets import IdMapSet_per_speaker
Anthony Larcher's avatar
debug    
Anthony Larcher committed
51
from .xsets import SpkSet
Anthony Larcher's avatar
Anthony Larcher committed
52
from .res_net import RawPreprocessor, ResBlockWFMS, ResBlock, PreResNet34, PreFastResNet34
Anthony Larcher's avatar
Anthony Larcher committed
53
from ..bosaris import IdMap
Anthony Larcher's avatar
Anthony Larcher committed
54
55
56
57
from ..bosaris import Key
from ..bosaris import Ndx
from ..bosaris.detplot import rocch
from ..bosaris.detplot import rocch2eer
Anthony Larcher's avatar
Anthony Larcher committed
58
from ..statserver import StatServer
Anthony Larcher's avatar
Anthony Larcher committed
59
from ..iv_scoring import cosine_scoring
Anthony Larcher's avatar
Anthony Larcher committed
60
from torch.utils.data import DataLoader
Anthony Larcher's avatar
Anthony Larcher committed
61
from sklearn.model_selection import train_test_split
Anthony Larcher's avatar
Anthony Larcher committed
62
from .sincnet import SincNet
Anthony Larcher's avatar
arcface    
Anthony Larcher committed
63
64
65
from .loss import ArcLinear
from .loss import ArcFace
from .loss import l2_norm
Anthony Larcher's avatar
Anthony Larcher committed
66
from .loss import ArcMarginProduct
Anthony Larcher's avatar
Anthony Larcher committed
67

Anthony Larcher's avatar
ddp    
Anthony Larcher committed
68

Anthony Larcher's avatar
Anthony Larcher committed
69
os.environ['MKL_THREADING_LAYER'] = 'GNU'
Anthony Larcher's avatar
debug    
Anthony Larcher committed
70

Anthony Larcher's avatar
Anthony Larcher committed
71
72
__license__ = "LGPL"
__author__ = "Anthony Larcher"
Anthony Larcher's avatar
v1.3.7    
Anthony Larcher committed
73
__copyright__ = "Copyright 2015-2021 Anthony Larcher"
Anthony Larcher's avatar
Anthony Larcher committed
74
75
76
77
__maintainer__ = "Anthony Larcher"
__email__ = "anthony.larcher@univ-lemans.fr"
__status__ = "Production"
__docformat__ = 'reS'
Anthony Larcher's avatar
Anthony Larcher committed
78
79


Anthony Larcher's avatar
Anthony Larcher committed
80
81
logging.basicConfig(format='%(asctime)s %(message)s')

Anthony Larcher's avatar
Anthony Larcher committed
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102

# Make PyTorch Deterministic
torch.manual_seed(0)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
numpy.random.seed(0)


class GuruMeditation (torch.autograd.detect_anomaly):
    
    def __init__(self):
        super(GuruMeditation, self).__init__()

    def __enter__(self):
        super(GuruMeditation, self).__enter__()
        return self

    def __exit__(self, type, value, trace):
        super(GuruMeditation, self).__exit__()
        if isinstance(value, RuntimeError):
            traceback.print_tb(trace)
Anthony Larcher's avatar
Anthony Larcher committed
103
            self.halt(str(value))
Anthony Larcher's avatar
Anthony Larcher committed
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127

    def halt(msg):
        print (msg)
        pdb.set_trace()


def select_n_random(data, labels, n=100):
    '''
    Selects n random datapoints and their corresponding labels from a dataset
    '''
    assert len(data) == len(labels)

    perm = torch.randperm(len(data))
    return data[perm][:n], labels[perm][:n]


def matplotlib_imshow(img, one_channel=False):
    if one_channel:
        img = img.mean(dim=0)
    img = img / 2 + 0.5     # unnormalize
    npimg = img.cpu().numpy()
    if one_channel:
        plt.imshow(npimg, cmap="Greys")
    else:
Anthony Larcher's avatar
Anthony Larcher committed
128
        plt.imshow(numpy.transpose(npimg, (1, 2, 0)))
Anthony Larcher's avatar
Anthony Larcher committed
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143

def speech_to_probs(model, speech):
    '''
    Generates predictions and corresponding probabilities from a trained
    network and a list of images
    '''
    output = model(speech)
    # convert output probabilities to predicted class
    _, preds_tensor = torch.max(output, 1)
    preds = numpy.squeeze(preds_tensor.cpu().numpy())
    return preds, [torch.nn.functional.softmax(el, dim=0)[i].item() for i, el in zip(preds, output)]


def plot_classes_preds(model, speech, labels):
    '''
Anthony Larcher's avatar
Anthony Larcher committed
144

Anthony Larcher's avatar
Anthony Larcher committed
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
    Generates matplotlib Figure using a trained network, along with images
    and labels from a batch, that shows the network's top prediction along
    with its probability, alongside the actual label, coloring this
    information based on whether the prediction was correct or not.
    Uses the "speech_to_probs" function.
    '''
    preds, probs = speech_to_probs(model, speech)
    # plot the images in the batch, along with predicted and true labels
    fig = plt.figure(figsize=(12, 48))
    for idx in numpy.arange(4):
         ax = fig.add_subplot(1, 4, idx+1, xticks=[], yticks=[])
         #matplotlib_imshow(speech[idx], one_channel=True)
         ax.set_title("{0}, {1:.1f}%\n(label: {2})".format(
                                                           preds[idx],
                                                           probs[idx] * 100.0,
                                                           labels[idx]),
                                                           color=("green" if preds[idx]==labels[idx].item() else "red"))
    return fig


Anthony Larcher's avatar
debug    
Anthony Larcher committed
165
166
def test_metrics(model,
                 device,
Anthony Larcher's avatar
Anthony Larcher committed
167
168
169
                 speaker_number,
                 num_thread,
                 mixed_precision):
Anthony Larcher's avatar
Anthony Larcher committed
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
    """Compute model metrics

    Args:
        model ([type]): [description]
        validation_loader ([type]): [description]
        device ([type]): [description]
        speaker_number ([type]): [description]
        model_archi ([type]): [description]

    Raises:
        NotImplementedError: [description]
        NotImplementedError: [description]

    Returns:
        [type]: [description]
    """
Anthony Larcher's avatar
debug    
Anthony Larcher committed
186
187
188
    idmap_test_filename = 'h5f/idmap_test.h5'
    ndx_test_filename = 'h5f/ndx_test.h5'
    key_test_filename = 'h5f/key_test.h5'
Anthony Larcher's avatar
debug    
Anthony Larcher committed
189
    data_root_name='/lium/corpus/base/voxceleb1/test/wav'
Anthony Larcher's avatar
debug    
Anthony Larcher committed
190
191

    transform_pipeline = dict()
Anthony Larcher's avatar
debug    
Anthony Larcher committed
192
193
194
195
196
197
198
199
200
201
    #mfcc_config = dict()
    #mfcc_config['nb_filters'] = 81
    #mfcc_config['nb_ceps'] = 80
    #mfcc_config['lowfreq'] = 133.333
    #mfcc_config['maxfreq'] = 6855.4976
    #mfcc_config['win_time'] = 0.025
    #mfcc_config['shift'] = 0.01
    #mfcc_config['n_fft'] = 2048
    #transform_pipeline['MFCC'] = mfcc_config
    #transform_pipeline['CMVN'] = {}
Anthony Larcher's avatar
debug    
Anthony Larcher committed
202
203
204
205
206
207

    xv_stat = extract_embeddings(idmap_name=idmap_test_filename,
                                 speaker_number=speaker_number,
                                 model_filename=model,
                                 data_root_name=data_root_name,
                                 device=device,
Anthony Larcher's avatar
Anthony Larcher committed
208
209
210
                                 transform_pipeline=transform_pipeline,
                                 num_thread=num_thread,
                                 mixed_precision=mixed_precision)
Anthony Larcher's avatar
debug    
Anthony Larcher committed
211
212
213
214
215
216
217
218
219
220
221
222

    scores = cosine_scoring(xv_stat,
                            xv_stat,
                            Ndx(ndx_test_filename),
                            wccn=None,
                            check_missing=True)

    tar, non = scores.get_tar_non(Key(key_test_filename))

    test_eer = eer(numpy.array(non).astype(numpy.double), numpy.array(tar).astype(numpy.double))

    return test_eer
Anthony Larcher's avatar
Anthony Larcher committed
223

Anthony Larcher's avatar
Anthony Larcher committed
224

225
def get_lr(optimizer):
Anthony Larcher's avatar
Anthony Larcher committed
226
227
228
229
230
    """

    :param optimizer:
    :return:
    """
231
232
233
234
    for param_group in optimizer.param_groups:
        return param_group['lr']


Anthony Larcher's avatar
Anthony Larcher committed
235
def save_checkpoint(state, is_best, filename='checkpoint.pth.tar', best_filename='model_best.pth.tar'):
Anthony Larcher's avatar
Anthony Larcher committed
236
237
238
239
240
241
242
243
    """

    :param state:
    :param is_best:
    :param filename:
    :param best_filename:
    :return:
    """
Anthony Larcher's avatar
Anthony Larcher committed
244
245
246
    torch.save(state, filename)
    if is_best:
        shutil.copyfile(filename, best_filename)
Anthony Larcher's avatar
Anthony Larcher committed
247

Anthony Larcher's avatar
Anthony Larcher committed
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
class MeanStdPooling(torch.nn.Module):
    """
    Mean and Standard deviation pooling
    """
    def __init__(self):
        """

        """
        super(MeanStdPooling, self).__init__()
        pass

    def forward(self, x):
        """

        :param x:
        :return:
        """
        mean = torch.mean(x, dim=2)
        std = torch.std(x, dim=2)
        return torch.cat([mean, std], dim=1)
Anthony Larcher's avatar
Anthony Larcher committed
268

Anthony Larcher's avatar
Anthony Larcher committed
269

Anthony Larcher's avatar
Anthony Larcher committed
270
271
272
273
274
275
276
277
278
279
280
class GruPooling(torch.nn.Module):
    """

    """
    def __init__(self, input_size, gru_node, nb_gru_layer):
        """

        :param input_size:
        :param gru_node:
        :param nb_gru_layer:
        """
Anthony Larcher's avatar
Anthony Larcher committed
281
        super(GruPooling, self).__init__()
Anthony Larcher's avatar
Anthony Larcher committed
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
        self.lrelu_keras = torch.nn.LeakyReLU(negative_slope = 0.3)
        self.bn_before_gru = torch.nn.BatchNorm1d(num_features = input_size)
        self.gru = torch.nn.GRU(input_size = input_size,
                          hidden_size = gru_node,
                          num_layers = nb_gru_layer,
                          batch_first = True)

    def forward(self, x):
        """

        :param x:
        :return:
        """
        x = self.bn_before_gru(x)
        x = self.lrelu_keras(x)
        x = x.permute(0, 2, 1)  #(batch, filt, time) >> (batch, time, filt)
        self.gru.flatten_parameters()
        x, _ = self.gru(x)
        x = x[:,-1,:]

        return x

Anthony Larcher's avatar
Anthony Larcher committed
304

Anthony Larcher's avatar
Anthony Larcher committed
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
class PreEmphasis(torch.nn.Module):

    def __init__(self, coef: float = 0.97):
        super().__init__()
        self.coef = coef
        # make kernel
        # In pytorch, the convolution operation uses cross-correlation. So, filter is flipped.
        self.register_buffer(
            'flipped_filter', torch.FloatTensor([-self.coef, 1.]).unsqueeze(0).unsqueeze(0)
        )

    def forward(self, input: torch.tensor) -> torch.tensor:
        assert len(input.size()) == 2, 'The number of dimensions of input tensor must be 2!'
        # reflect padding to match lengths of in/out
        input = input.unsqueeze(1)
        input = torch.nn.functional.pad(input, (1, 0), 'reflect')
        return torch.nn.functional.conv1d(input, self.flipped_filter).squeeze(1)


class MfccFrontEnd(torch.nn.Module):
    """

    """

    def __init__(self,
                 pre_emphasis=0.97,
                 sample_rate=16000,
                 n_fft=2048,
                 f_min=133.333,
                 f_max=6855.4976,
                 win_length=1024,
                 window_fn=torch.hann_window,
                 hop_length=512,
                 power=2.0,
                 n_mels=100,
                 n_mfcc=80):

        super(MfccFrontEnd, self).__init__()

        self.pre_emphasis = pre_emphasis
        self.sample_rate = sample_rate
        self.n_fft = n_fft
        self.f_min = f_min
        self.f_max = f_max
        self.win_length = win_length
        self.window_fn=window_fn
        self.hop_length = hop_length
        self.power=power
        self.window_fn = window_fn
        self.n_mels = n_mels
        self.n_mfcc = n_mfcc

        self.PreEmphasis = PreEmphasis(self.pre_emphasis)

Anthony Larcher's avatar
Anthony Larcher committed
359
360
361
362
363
364
365
366
        self.melkwargs = {"n_fft":self.n_fft,
                          "f_min":self.f_min,
                          "f_max":self.f_max,
                          "win_length":self.win_length,
                          "window_fn":self.window_fn,
                          "hop_length":self.hop_length,
                          "power":self.power,
                          "n_mels":self.n_mels}
Anthony Larcher's avatar
Anthony Larcher committed
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389

        self.MFCC = torchaudio.transforms.MFCC(
            sample_rate=self.sample_rate,
            n_mfcc=self.n_mfcc,
            dct_type=2,
            log_mels=True,
            melkwargs=self.melkwargs)

        self.CMVN = torch.nn.InstanceNorm1d(self.n_mfcc)

    def forward(self, x):
        """

        :param x:
        :return:
        """
        with torch.no_grad():
            with torch.cuda.amp.autocast(enabled=False):
                mfcc = self.PreEmphasis(x)
                mfcc = self.MFCC(mfcc)
                mfcc = self.CMVN(mfcc)
        return mfcc

Anthony Larcher's avatar
Anthony Larcher committed
390
class Xtractor(torch.nn.Module):
391
392
393
    """
    Class that defines an x-vector extractor based on 5 convolutional layers and a mean standard deviation pooling
    """
Anthony Larcher's avatar
minor    
Anthony Larcher committed
394

Anthony Larcher's avatar
Anthony Larcher committed
395
396
397
    def __init__(self,
                 speaker_number,
                 model_archi="xvector",
Anthony Larcher's avatar
Anthony Larcher committed
398
                 loss=None,
Anthony Larcher's avatar
Anthony Larcher committed
399
                 norm_embedding=False,
Anthony Larcher's avatar
debug    
Anthony Larcher committed
400
401
                 aam_margin=0.2,
                 aam_s=30):
Anthony Larcher's avatar
Anthony Larcher committed
402
403
        """
        If config is None, default architecture is created
Anthony Larcher's avatar
Anthony Larcher committed
404
        :param model_archi:
Anthony Larcher's avatar
Anthony Larcher committed
405
        """
Anthony Larcher's avatar
Anthony Larcher committed
406
        super(Xtractor, self).__init__()
Anthony Larcher's avatar
Anthony Larcher committed
407
        self.speaker_number = speaker_number
Anthony Larcher's avatar
Anthony Larcher committed
408
        self.feature_size = None
Anthony Larcher's avatar
Anthony Larcher committed
409
        self.norm_embedding = norm_embedding
Anthony Larcher's avatar
Anthony Larcher committed
410

Anthony Larcher's avatar
Anthony Larcher committed
411
412
        print(f"Speaker number : {self.speaker_number}")

Anthony Larcher's avatar
Anthony Larcher committed
413
        if model_archi == "xvector":
Anthony Larcher's avatar
Anthony Larcher committed
414

Anthony Larcher's avatar
Anthony Larcher committed
415
416
            self.input_nbdim = 2

Anthony Larcher's avatar
Anthony Larcher committed
417
418
419
420
421
            if loss not in ["cce", 'aam']:
                raise NotImplementedError(f"The valid loss are for now cce and aam ")
            else:
                self.loss = loss

Anthony Larcher's avatar
Anthony Larcher committed
422
423
            self.activation = torch.nn.LeakyReLU(0.2)

Anthony Larcher's avatar
Anthony Larcher committed
424
            self.preprocessor = MfccFrontEnd()
Anthony Larcher's avatar
debug    
Anthony Larcher committed
425
            self.feature_size = self.preprocessor.n_mfcc
Anthony Larcher's avatar
Anthony Larcher committed
426

Anthony Larcher's avatar
xv    
Anthony Larcher committed
427
            self.sequence_network = torch.nn.Sequential(OrderedDict([
Anthony Larcher's avatar
Anthony Larcher committed
428
                ("conv1", torch.nn.Conv1d(self.feature_size, 512, 5, dilation=1)),
Anthony Larcher's avatar
Anthony Larcher committed
429
                ("activation1", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
430
                ("batch_norm1", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
431
432
                ("conv2", torch.nn.Conv1d(512, 512, 3, dilation=2)),
                ("activation2", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
433
                ("batch_norm2", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
434
435
                ("conv3", torch.nn.Conv1d(512, 512, 3, dilation=3)),
                ("activation3", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
436
                ("batch_norm3", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
437
                ("conv4", torch.nn.Conv1d(512, 512, 1)),
Anthony Larcher's avatar
Anthony Larcher committed
438
                ("activation4", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
439
                ("batch_norm4", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
440
                ("conv5", torch.nn.Conv1d(512, 1536, 1)),
Anthony Larcher's avatar
Anthony Larcher committed
441
                ("activation5", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
442
                ("batch_norm5", torch.nn.BatchNorm1d(1536))
Anthony Larcher's avatar
Anthony Larcher committed
443
444
            ]))

Anthony Larcher's avatar
pooling    
Anthony Larcher committed
445
            self.stat_pooling = MeanStdPooling()
Anthony Larcher's avatar
Anthony Larcher committed
446
            self.stat_pooling_weight_decay = 0
Anthony Larcher's avatar
xv    
Anthony Larcher committed
447
            self.before_speaker_embedding = torch.nn.Sequential(OrderedDict([
Anthony Larcher's avatar
Anthony Larcher committed
448
                ("linear6", torch.nn.Linear(3072, 512))
Anthony Larcher's avatar
Anthony Larcher committed
449
450
            ]))

451
452
            self.embedding_size = 512

Anthony Larcher's avatar
Anthony Larcher committed
453
            if self.loss == "aam":
Anthony Larcher's avatar
Anthony Larcher committed
454
455
456
457
458
                self.after_speaker_embedding = ArcMarginProduct(512,
                                                                int(self.speaker_number),
                                                                s=64,
                                                                m=0.2,
                                                                easy_margin=True)
Anthony Larcher's avatar
Anthony Larcher committed
459
460
461
            elif self.loss == "cce":
                self.after_speaker_embedding = torch.nn.Sequential(OrderedDict([
                    ("activation6", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
462
                    ("batch_norm6", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
463
464
465
                    ("dropout6", torch.nn.Dropout(p=0.05)),
                    ("linear7", torch.nn.Linear(512, 512)),
                    ("activation7", torch.nn.LeakyReLU(0.2)),
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
466
                    ("batch_norm7", torch.nn.BatchNorm1d(512)),
Anthony Larcher's avatar
Anthony Larcher committed
467
468
                    ("linear8", torch.nn.Linear(512, int(self.speaker_number)))
                ]))
Anthony Larcher's avatar
Anthony Larcher committed
469

Anthony Larcher's avatar
debug    
Anthony Larcher committed
470
471
472
            self.sequence_network_weight_decay = 0.0002
            self.before_speaker_embedding_weight_decay = 0.002
            self.after_speaker_embedding_weight_decay = 0.002
473
            self.embedding_size = 512
Anthony Larcher's avatar
Anthony Larcher committed
474

Anthony Larcher's avatar
Anthony Larcher committed
475
        elif model_archi == "resnet34":
Anthony Larcher's avatar
debug    
Anthony Larcher committed
476

Anthony Larcher's avatar
Anthony Larcher committed
477
            self.preprocessor = MfccFrontEnd()
Anthony Larcher's avatar
Anthony Larcher committed
478
479
480
481
482
483
484
485
            self.sequence_network = PreResNet34()

            self.before_speaker_embedding = torch.nn.Linear(in_features = 5120,
                                                            out_features = 256)

            self.stat_pooling = MeanStdPooling()
            self.stat_pooling_weight_decay = 0

486
            self.embedding_size = 256
Anthony Larcher's avatar
Anthony Larcher committed
487

488
            self.loss = "aam"
Anthony Larcher's avatar
Anthony Larcher committed
489
490
491
            self.after_speaker_embedding = ArcMarginProduct(256,
                                                            int(self.speaker_number),
                                                            s = 30.0,
Anthony Larcher's avatar
debug    
Anthony Larcher committed
492
                                                            m = 0.20,
Anthony Larcher's avatar
Anthony Larcher committed
493
                                                            easy_margin = True)
Anthony Larcher's avatar
Anthony Larcher committed
494
495
496
497
498
499
500

            self.preprocessor_weight_decay = 0.000
            self.sequence_network_weight_decay = 0.000
            self.stat_pooling_weight_decay = 0.000
            self.before_speaker_embedding_weight_decay = 0.00
            self.after_speaker_embedding_weight_decay = 0.00

Anthony Larcher's avatar
Anthony Larcher committed
501
        elif model_archi == "fastresnet34":
Anthony Larcher's avatar
debug    
Anthony Larcher committed
502

Anthony Larcher's avatar
Anthony Larcher committed
503
            self.preprocessor = MfccFrontEnd()
Anthony Larcher's avatar
Anthony Larcher committed
504
505
            self.sequence_network = PreFastResNet34()

Anthony Larcher's avatar
Anthony Larcher committed
506
            self.before_speaker_embedding = torch.nn.Linear(in_features = 2560,
Anthony Larcher's avatar
Anthony Larcher committed
507
508
509
510
511
512
513
514
515
516
517
518
                                                            out_features = 256)

            self.stat_pooling = MeanStdPooling()
            self.stat_pooling_weight_decay = 0

            self.embedding_size = 256

            self.loss = "aam"
            self.after_speaker_embedding = ArcMarginProduct(256,
                                                            int(self.speaker_number),
                                                            s = 30.0,
                                                            m = 0.20,
Anthony Larcher's avatar
debug    
Anthony Larcher committed
519
                                                            easy_margin = False)
Anthony Larcher's avatar
Anthony Larcher committed
520
521
522
523
524
525

            self.preprocessor_weight_decay = 0.000
            self.sequence_network_weight_decay = 0.000
            self.stat_pooling_weight_decay = 0.000
            self.before_speaker_embedding_weight_decay = 0.00
            self.after_speaker_embedding_weight_decay = 0.00
Anthony Larcher's avatar
Anthony Larcher committed
526

Anthony Larcher's avatar
Anthony Larcher committed
527
        elif model_archi == "rawnet2":
Anthony Larcher's avatar
Anthony Larcher committed
528
529
530
531
532
533

            if loss not in ["cce", 'aam']:
                raise NotImplementedError(f"The valid loss are for now cce and aam ")
            else:
                self.loss = loss

Anthony Larcher's avatar
Anthony Larcher committed
534
535
            self.input_nbdim = 2

Anthony Larcher's avatar
Anthony Larcher committed
536
537
538
            filts = [128, [128, 128], [128, 256], [256, 256]]
            self.norm_embedding = True

Anthony Larcher's avatar
Anthony Larcher committed
539
            self.preprocessor = RawPreprocessor(nb_samp=48000,
Anthony Larcher's avatar
Anthony Larcher committed
540
                                                in_channels=1,
Anthony Larcher's avatar
Anthony Larcher committed
541
542
                                                out_channels=filts[0],
                                                kernel_size=3)
Anthony Larcher's avatar
Anthony Larcher committed
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559

            self.sequence_network = torch.nn.Sequential(OrderedDict([
                ("block0", ResBlockWFMS(nb_filts=filts[1], first=True)),
                ("block1", ResBlockWFMS(nb_filts=filts[1])),
                ("block2", ResBlockWFMS(nb_filts=filts[2])),
                ("block3", ResBlockWFMS(nb_filts=[filts[2][1], filts[2][1]])),
                ("block4", ResBlockWFMS(nb_filts=[filts[2][1], filts[2][1]])),
                ("block5", ResBlockWFMS(nb_filts=[filts[2][1], filts[2][1]]))
            ]))

            self.stat_pooling = GruPooling(input_size=filts[2][-1],
                                           gru_node=1024,
                                           nb_gru_layer=1)

            self.before_speaker_embedding = torch.nn.Linear(in_features = 1024,
                                                            out_features = 1024)

Anthony Larcher's avatar
Anthony Larcher committed
560
561
562
563
564
            if self.loss == "aam":
                if loss == 'aam':
                    self.after_speaker_embedding = ArcLinear(1024,
                                                             int(self.speaker_number),
                                                             margin=aam_margin, s=aam_s)
Anthony Larcher's avatar
Anthony Larcher committed
565
            elif self.loss == "cce":
Anthony Larcher's avatar
Anthony Larcher committed
566
567
568
                self.after_speaker_embedding = torch.nn.Linear(in_features = 1024,
                                                               out_features = int(self.speaker_number),
                                                               bias = True)
Anthony Larcher's avatar
Anthony Larcher committed
569

Anthony Larcher's avatar
rawnet2    
Anthony Larcher committed
570
571
572
573
574
575
            self.preprocessor_weight_decay = 0.000
            self.sequence_network_weight_decay = 0.000
            self.stat_pooling_weight_decay = 0.000
            self.before_speaker_embedding_weight_decay = 0.00
            self.after_speaker_embedding_weight_decay = 0.00

Anthony Larcher's avatar
Anthony Larcher committed
576
        else:
Anthony Larcher's avatar
Anthony Larcher committed
577
578
            is_first_resblock = True

Anthony Larcher's avatar
fix API    
Anthony Larcher committed
579
580
581
582
583
584
            if isinstance(model_archi, dict):
                cfg = model_archi
            else:
                # Load Yaml configuration
                with open(model_archi, 'r') as fh:
                    cfg = yaml.load(fh, Loader=yaml.FullLoader)
Anthony Larcher's avatar
Anthony Larcher committed
585

Anthony Larcher's avatar
fix API    
Anthony Larcher committed
586
            self.loss = cfg["training"]["loss"]
Anthony Larcher's avatar
Anthony Larcher committed
587
            if self.loss == "aam":
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
588
589
                self.aam_margin = cfg["training"]["aam_margin"]
                self.aam_s = cfg["training"]["aam_s"]
Anthony Larcher's avatar
Anthony Larcher committed
590

Anthony Larcher's avatar
Anthony Larcher committed
591
592
593
            """
            Prepare Preprocessor
            """
Anthony Larcher's avatar
Anthony Larcher committed
594
            self.preprocessor = None
Anthony Larcher's avatar
Anthony Larcher committed
595
596
            if "preprocessor" in cfg:
                if cfg['preprocessor']["type"] == "sincnet":
Anthony Larcher's avatar
Anthony Larcher committed
597
                    self.preprocessor = SincNet(
Anthony Larcher's avatar
Anthony Larcher committed
598
599
600
601
602
603
604
605
606
607
608
609
                        waveform_normalize=cfg['preprocessor']["waveform_normalize"],
                        sample_rate=cfg['preprocessor']["sample_rate"],
                        min_low_hz=cfg['preprocessor']["min_low_hz"],
                        min_band_hz=cfg['preprocessor']["min_band_hz"],
                        out_channels=cfg['preprocessor']["out_channels"],
                        kernel_size=cfg['preprocessor']["kernel_size"],
                        stride=cfg['preprocessor']["stride"],
                        max_pool=cfg['preprocessor']["max_pool"],
                        instance_normalize=cfg['preprocessor']["instance_normalize"],
                        activation=cfg['preprocessor']["activation"],
                        dropout=cfg['preprocessor']["dropout"]
                    )
Anthony Larcher's avatar
Anthony Larcher committed
610
                    self.feature_size = self.preprocessor.dimension
Anthony Larcher's avatar
Anthony Larcher committed
611
                elif cfg['preprocessor']["type"] == "rawnet2":
Anthony Larcher's avatar
Anthony Larcher committed
612
                    self.preprocessor = RawPreprocessor(nb_samp=int(cfg['preprocessor']["sampling_frequency"] * cfg['preprocessor']["duration"]),
Anthony Larcher's avatar
Anthony Larcher committed
613
                                                        in_channels=1,
Anthony Larcher's avatar
Anthony Larcher committed
614
615
616
617
618
619
                                                        out_channels=cfg["feature_size"],
                                                        kernel_size=cfg['preprocessor']["kernel_size"],
                                                        stride=cfg['preprocessor']["stride"],
                                                        padding=cfg['preprocessor']["padding"],
                                                        dilation=cfg['preprocessor']["dilation"])
                    self.feature_size = cfg["feature_size"]
Anthony Larcher's avatar
Anthony Larcher committed
620
                self.preprocessor_weight_decay = 0.000
Anthony Larcher's avatar
Anthony Larcher committed
621
622

            """
Anthony Larcher's avatar
minor    
Anthony Larcher committed
623
            Prepare sequence network
Anthony Larcher's avatar
Anthony Larcher committed
624
            """
Anthony Larcher's avatar
Anthony Larcher committed
625
            # Get Feature size
Anthony Larcher's avatar
Anthony Larcher committed
626
627
628
            if self.feature_size is None:
                self.feature_size = cfg["feature_size"]

Anthony Larcher's avatar
Anthony Larcher committed
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
            input_size = self.feature_size

            # Get activation function
            if cfg["activation"] == 'LeakyReLU':
                self.activation = torch.nn.LeakyReLU(0.2)
            elif cfg["activation"] == 'PReLU':
                self.activation = torch.nn.PReLU()
            elif cfg["activation"] == 'ReLU6':
                self.activation = torch.nn.ReLU6()
            else:
                self.activation = torch.nn.ReLU()

            # Create sequential object for the first part of the network
            segmental_layers = []
            for k in cfg["segmental"].keys():
Anthony Larcher's avatar
Anthony Larcher committed
644
645
646
647
648
                if k.startswith("lin"):
                    segmental_layers.append((k, torch.nn.Linear(input_size,
                                                                cfg["segmental"][k]["output"])))
                    input_size = cfg["segmental"][k]["output"]

Anthony Larcher's avatar
Anthony Larcher committed
649
650
651
652
653
                elif k.startswith("conv2D"):
                    segmental_layers.append((k, torch.nn.Conv2d(in_channels=1,
                                                                out_channels=entry_conv_out_channels,
                                                                kernel_size=entry_conv_kernel_size,
                                                                padding=3,
Anthony Larcher's avatar
Anthony Larcher committed
654
                                                                stride=1)))
Anthony Larcher's avatar
Anthony Larcher committed
655

Anthony Larcher's avatar
Anthony Larcher committed
656
                elif k.startswith("conv"):
Anthony Larcher's avatar
Anthony Larcher committed
657
                    segmental_layers.append((k, torch.nn.Conv1d(input_size,
Anthony Larcher's avatar
Anthony Larcher committed
658
                                                                cfg["segmental"][k]["output_channels"],
Anthony Larcher's avatar
Anthony Larcher committed
659
660
                                                                kernel_size=cfg["segmental"][k]["kernel_size"],
                                                                dilation=cfg["segmental"][k]["dilation"])))
Anthony Larcher's avatar
Anthony Larcher committed
661
662
                    input_size = cfg["segmental"][k]["output_channels"]

Anthony Larcher's avatar
Anthony Larcher committed
663
664
                elif k.startswith("ctrans"):
                    segmental_layers.append((k, torch.nn.ConvTranspose1d(input_size,
Anthony Larcher's avatar
minor    
Anthony Larcher committed
665
                                                                         cfg["segmental"][k]["output_channels"],
Anthony Larcher's avatar
Anthony Larcher committed
666
667
                                                                         kernel_size=cfg["segmental"][k]["kernel_size"],
                                                                         dilation=cfg["segmental"][k]["dilation"])))
Anthony Larcher's avatar
Anthony Larcher committed
668
669
670
                elif k.startswith("activation"):
                    segmental_layers.append((k, self.activation))

Anthony Larcher's avatar
Anthony Larcher committed
671
                elif k.startswith('batch_norm'):
Anthony Larcher's avatar
Anthony Larcher committed
672
673
                    segmental_layers.append((k, torch.nn.BatchNorm1d(input_size)))

Anthony Larcher's avatar
Anthony Larcher committed
674
675
676
677
678
679
                elif k.startswith('resblock'):
                    segmental_layers.append((ResBlock(cfg["segmental"][k]["input_channel"],
                                                      cfg["segmental"][k]["output_channel"],
                                                      is_first_resblock)))
                    is_first_resblock = False

Anthony Larcher's avatar
Anthony Larcher committed
680
            self.sequence_network = torch.nn.Sequential(OrderedDict(segmental_layers))
Anthony Larcher's avatar
Anthony Larcher committed
681
            self.sequence_network_weight_decay = cfg["segmental"]["weight_decay"]
Anthony Larcher's avatar
Anthony Larcher committed
682

Anthony Larcher's avatar
pooling    
Anthony Larcher committed
683
684
685
686
            """
            Pooling
            """
            self.stat_pooling = MeanStdPooling()
Anthony Larcher's avatar
Anthony Larcher committed
687
            tmp_input_size = input_size * 2
Anthony Larcher's avatar
pooling    
Anthony Larcher committed
688
689
690
691
            if cfg["stat_pooling"]["type"] == "GRU":
                self.stat_pooling = GruPooling(input_size=cfg["stat_pooling"]["input_size"],
                                               gru_node=cfg["stat_pooling"]["gru_node"],
                                               nb_gru_layer=cfg["stat_pooling"]["nb_gru_layer"])
Anthony Larcher's avatar
debug    
Anthony Larcher committed
692
                tmp_input_size = cfg["stat_pooling"]["gru_node"]
Anthony Larcher's avatar
pooling    
Anthony Larcher committed
693

Anthony Larcher's avatar
Anthony Larcher committed
694
695
            self.stat_pooling_weight_decay = cfg["stat_pooling"]["weight_decay"]

Anthony Larcher's avatar
Anthony Larcher committed
696
            """
Anthony Larcher's avatar
Anthony Larcher committed
697
            Prepare last part of the network (after pooling)
Anthony Larcher's avatar
Anthony Larcher committed
698
            """
Anthony Larcher's avatar
Anthony Larcher committed
699
            # Create sequential object for the second part of the network
Anthony Larcher's avatar
Anthony Larcher committed
700
            input_size = tmp_input_size
Anthony Larcher's avatar
xv    
Anthony Larcher committed
701
702
            before_embedding_layers = []
            for k in cfg["before_embedding"].keys():
Anthony Larcher's avatar
Anthony Larcher committed
703
                if k.startswith("lin"):
Anthony Larcher's avatar
Anthony Larcher committed
704
705
                    if cfg["before_embedding"][k]["output"] == "speaker_number":
                        before_embedding_layers.append((k, torch.nn.Linear(input_size, self.speaker_number)))
Anthony Larcher's avatar
Anthony Larcher committed
706
                    else:
Anthony Larcher's avatar
Anthony Larcher committed
707
708
                        before_embedding_layers.append((k, torch.nn.Linear(input_size,
                                                                           cfg["before_embedding"][k]["output"])))
Anthony Larcher's avatar
Anthony Larcher committed
709
                        input_size = cfg["before_embedding"][k]["output"]
Anthony Larcher's avatar
Anthony Larcher committed
710
711

                elif k.startswith("activation"):
Anthony Larcher's avatar
xv    
Anthony Larcher committed
712
                    before_embedding_layers.append((k, self.activation))
Anthony Larcher's avatar
Anthony Larcher committed
713

Anthony Larcher's avatar
Anthony Larcher committed
714
                elif k.startswith('batch_norm'):
Anthony Larcher's avatar
xv    
Anthony Larcher committed
715
                    before_embedding_layers.append((k, torch.nn.BatchNorm1d(input_size)))
Anthony Larcher's avatar
Anthony Larcher committed
716
717

                elif k.startswith('dropout'):
Anthony Larcher's avatar
Anthony Larcher committed
718
                    before_embedding_layers.append((k, torch.nn.Dropout(p=cfg["before_embedding"][k])))
Anthony Larcher's avatar
xv    
Anthony Larcher committed
719

Anthony Larcher's avatar
Anthony Larcher committed
720
            self.embedding_size = input_size
Anthony Larcher's avatar
Anthony Larcher committed
721
            self.before_speaker_embedding = torch.nn.Sequential(OrderedDict(before_embedding_layers))
Anthony Larcher's avatar
Anthony Larcher committed
722
            self.before_speaker_embedding_weight_decay = cfg["before_embedding"]["weight_decay"]
Anthony Larcher's avatar
xv    
Anthony Larcher committed
723

Anthony Larcher's avatar
Anthony Larcher committed
724
            # if loss_criteria is "cce"
Anthony Larcher's avatar
xv    
Anthony Larcher committed
725
            # Create sequential object for the second part of the network
Anthony Larcher's avatar
Anthony Larcher committed
726
727
728
729
730
731
732
733
            if self.loss == "cce":
                after_embedding_layers = []
                for k in cfg["after_embedding"].keys():
                    if k.startswith("lin"):
                        if cfg["after_embedding"][k]["output"] == "speaker_number":
                            after_embedding_layers.append((k, torch.nn.Linear(input_size, self.speaker_number)))
                        else:
                            after_embedding_layers.append((k, torch.nn.Linear(input_size,
Anthony Larcher's avatar
Anthony Larcher committed
734
                                                                          cfg["after_embedding"][k]["output"])))
Anthony Larcher's avatar
Anthony Larcher committed
735
                            input_size = cfg["after_embedding"][k]["output"]
Anthony Larcher's avatar
xv    
Anthony Larcher committed
736

Anthony Larcher's avatar
Anthony Larcher committed
737
738
739
740
741
                    elif k.startswith('arc'):
                        after_embedding_layers.append((k, ArcLinear(input_size,
                                                                    self.speaker_number,
                                                                    margin=self.aam_margin,
                                                                    s=self.aam_s)))
Anthony Larcher's avatar
Anthony Larcher committed
742

Anthony Larcher's avatar
Anthony Larcher committed
743
744
                    elif k.startswith("activation"):
                        after_embedding_layers.append((k, self.activation))
Anthony Larcher's avatar
xv    
Anthony Larcher committed
745

Anthony Larcher's avatar
Anthony Larcher committed
746
747
                    elif k.startswith('batch_norm'):
                        after_embedding_layers.append((k, torch.nn.BatchNorm1d(input_size)))
Anthony Larcher's avatar
xv    
Anthony Larcher committed
748

Anthony Larcher's avatar
Anthony Larcher committed
749
750
751
752
753
754
                    elif k.startswith('dropout'):
                        after_embedding_layers.append((k, torch.nn.Dropout(p=cfg["after_embedding"][k])))

                self.after_speaker_embedding = torch.nn.Sequential(OrderedDict(after_embedding_layers))

            elif self.loss == "aam":
Anthony Larcher's avatar
Anthony Larcher committed
755
756
757
758
759
760
761
                self.norm_embedding = True
                self.after_speaker_embedding = ArcMarginProduct(input_size,
                                                                int(self.speaker_number),
                                                                s=64,
                                                                m=0.2,
                                                                easy_margin=True)

Anthony Larcher's avatar
arcface    
Anthony Larcher committed
762
763
764
765
                #self.after_speaker_embedding = ArcLinear(input_size,
                #                                         self.speaker_number,
                #                                         margin=self.aam_margin,
                #                                         s=self.aam_s)
Anthony Larcher's avatar
Anthony Larcher committed
766
767
768
769
                #self.after_speaker_embedding = ArcFace(embedding_size=input_size,
                #                                       classnum=self.speaker_number,
                #                                       s=64.,
                #                                       m=0.5)
Anthony Larcher's avatar
Anthony Larcher committed
770

Anthony Larcher's avatar
Anthony Larcher committed
771
            self.after_speaker_embedding_weight_decay = cfg["after_embedding"]["weight_decay"]
Anthony Larcher's avatar
Anthony Larcher committed
772

Anthony Larcher's avatar
Anthony Larcher committed
773

774
    def forward(self, x, is_eval=False, target=None, extract_after_pooling=False):
775
776
777
        """

        :param x:
778
        :param is_eval: False for training
779
780
        :return:
        """
Anthony Larcher's avatar
Anthony Larcher committed
781
782
783
        if self.preprocessor is not None:
            x = self.preprocessor(x)

Anthony Larcher's avatar
Anthony Larcher committed
784
        x = self.sequence_network(x)
785

Anthony Larcher's avatar
Anthony Larcher committed
786
        # Mean and Standard deviation pooling
Anthony Larcher's avatar
pooling    
Anthony Larcher committed
787
        x = self.stat_pooling(x)
Anthony Larcher's avatar
Anthony Larcher committed
788

789
790
791
        if extract_after_pooling:
            return x

Anthony Larcher's avatar
Anthony Larcher committed
792
        x = self.before_speaker_embedding(x)
793

Anthony Larcher's avatar
Anthony Larcher committed
794
        if self.norm_embedding:
Anthony Larcher's avatar
arcface    
Anthony Larcher committed
795
            #x_norm = x.norm(p=2,dim=1, keepdim=True) / 10. # Why  10. ?
Anthony Larcher's avatar
arcface    
Anthony Larcher committed
796
797
798
            #x_norm = torch.linalg.norm(x, ord=2, dim=1, keepdim=True, out=None, dtype=None)
            #x = torch.div(x, x_norm)
            x = l2_norm(x)
Anthony Larcher's avatar
Anthony Larcher committed
799

Anthony Larcher's avatar
Anthony Larcher committed
800
        if self.loss == "cce":
Anthony Larcher's avatar
Anthony Larcher committed
801
            if is_eval:
Anthony Larcher's avatar
debug    
Anthony Larcher committed
802
                return x
Anthony Larcher's avatar
debug    
Anthony Larcher committed
803
            else:
Anthony Larcher's avatar
debug    
Anthony Larcher committed
804
                return self.after_speaker_embedding(x), x
Anthony Larcher's avatar
Anthony Larcher committed
805

Anthony Larcher's avatar
Anthony Larcher committed
806
        elif self.loss == "aam":
Anthony Larcher's avatar
debug    
Anthony Larcher committed
807
808
            if is_eval:
                x = torch.nn.functional.normalize(x, dim=1)
Anthony Larcher's avatar
debug    
Anthony Larcher committed
809
            else:
Anthony Larcher's avatar
debug    
Anthony Larcher committed
810
                x = self.after_speaker_embedding(torch.nn.functional.normalize(x, dim=1), target=target), torch.nn.functional.normalize(x, dim=1)
Anthony Larcher's avatar
Anthony Larcher committed
811

Anthony Larcher's avatar
Anthony Larcher committed
812
        return x
Anthony Larcher's avatar
Anthony Larcher committed
813

814
815
    def context_size(self):
        context = 1
Anthony Larcher's avatar
Anthony Larcher committed
816
817
818
819
820
821
822
823
824
        if isinstance(self, Xtractor):
            for name, module in self.sequence_network.named_modules():
                if name.startswith("conv"):
                    context += module.dilation[0] * (module.kernel_size[0] - 1)
        else:
            for name, module in self.module.sequence_network.named_modules():
                if name.startswith("conv"):
                    context += module.dilation[0] * (module.kernel_size[0] - 1)
        return context
Anthony Larcher's avatar
Anthony Larcher committed
825

Anthony Larcher's avatar
Anthony Larcher committed
826
def xtrain(speaker_number,
Anthony Larcher's avatar
Anthony Larcher committed
827
           dataset_yaml,
Anthony Larcher's avatar
Anthony Larcher committed
828
829
           epochs=None,
           lr=None,
Anthony Larcher's avatar
Anthony Larcher committed
830
           model_yaml=None,
Anthony Larcher's avatar
Anthony Larcher committed
831
           model_name=None,
Anthony Larcher's avatar
Anthony Larcher committed
832
833
834
835
           loss=None,
           aam_margin=None,
           aam_s=None,
           patience=None,
Anthony Larcher's avatar
Anthony Larcher committed
836
           tmp_model_name=None,
Anthony Larcher's avatar
minor    
Anthony Larcher committed
837
           best_model_name=None,
Anthony Larcher's avatar
Anthony Larcher committed
838
           multi_gpu=True,
Anthony Larcher's avatar
Anthony Larcher committed
839
           mixed_precision=False,
840
           clipping=False,
Anthony Larcher's avatar
Anthony Larcher committed
841
           opt=None,
Anthony Larcher's avatar
Anthony Larcher committed
842
843
           reset_parts=[],
           freeze_parts=[],
Anthony Larcher's avatar
Anthony Larcher committed
844
845
846
847
           num_thread=None,
           write_batches_to_disk=False,
           load_batches_from_disk=False,
           tmp_batch_dir=None):
848
849
    """

Anthony Larcher's avatar
Anthony Larcher committed
850
851
852
853
854
855
    :param speaker_number:
    :param dataset_yaml:
    :param epochs:
    :param lr:
    :param model_yaml:
    :param model_name:
Anthony Larcher's avatar
Anthony Larcher committed
856
857
858
859
    :param loss:
    :param aam_margin:
    :param aam_s:
    :param patience:
Anthony Larcher's avatar
Anthony Larcher committed
860
861
862
863
    :param tmp_model_name:
    :param best_model_name:
    :param multi_gpu:
    :param clipping:
Anthony Larcher's avatar
Anthony Larcher committed
864
865
866
    :param opt:
    :param reset_parts:
    :param freeze_parts:
Anthony Larcher's avatar
Anthony Larcher committed
867
    :param num_thread:
868
869
    :return:
    """
Anthony Larcher's avatar
Anthony Larcher committed
870
871
872
873
    # Add for tensorboard
    # Writer will output to ./runs/ directory by default
    #writer = SummaryWriter("runs/xvectors_experiments_2")
    writer = None
Anthony Larcher's avatar
Anthony Larcher committed
874

Anthony Larcher's avatar
Anthony Larcher committed
875
876
877
    if write_batches_to_disk:
        load_batches_from_disk = True

Anthony Larcher's avatar
Anthony Larcher committed
878
    if num_thread is None:
Anthony Larcher's avatar
debug    
Anthony Larcher committed
879
880
        import multiprocessing

Anthony Larcher's avatar
Anthony Larcher committed
881
882
        num_thread = multiprocessing.cpu_count()

Anthony Larcher's avatar
Anthony Larcher committed
883
    logging.critical(f"Use {num_thread} cpus")
Anthony Larcher's avatar
Anthony Larcher committed
884
    logging.critical(f"Start process at {time.strftime('%H:%M:%S', time.localtime())}")
885

Anthony Larcher's avatar
Anthony Larcher committed
886
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
Anthony Larcher's avatar
Anthony Larcher committed
887
    # Start from scratch
Anthony Larcher's avatar
Anthony Larcher committed
888
    if model_name is None and model_yaml in ["xvector", "rawnet2", "resnet34", "fastresnet34"]:
Anthony Larcher's avatar
Anthony Larcher committed
889
        # Initialize a first model
Anthony Larcher's avatar
rawnet2    
Anthony Larcher committed
890
        if model_yaml == "xvector":
Anthony Larcher's avatar
Anthony Larcher committed
891
            model = Xtractor(speaker_number, "xvector", loss=loss)
Anthony Larcher's avatar
rawnet2    
Anthony Larcher committed
892
        elif model_yaml == "rawnet2":
Anthony Larcher's avatar
pooling    
Anthony Larcher committed
893
            model = Xtractor(speaker_number, "rawnet2")
894
895
        elif model_yaml == "resnet34":
            model = Xtractor(speaker_number, "resnet34")
Anthony Larcher's avatar
Anthony Larcher committed
896
897
        elif model_yaml == "fastresnet34":
            model = Xtractor(speaker_number, "fastresnet34")
Anthony Larcher's avatar
Anthony Larcher committed
898
        model_archi = model_yaml
Anthony Larcher's avatar
Anthony Larcher committed
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
    else:
        with open(model_yaml, 'r') as fh:
            model_archi = yaml.load(fh, Loader=yaml.FullLoader)
            if epochs is None:
                epochs = model_archi["training"]["epochs"]
            if patience is None:
                patience = model_archi["training"]["patience"]
            if opt is None:
                opt = model_archi["training"]["opt"]
            if lr is None:
                lr = model_archi["training"]["lr"]
            if loss is None:
                loss = model_archi["training"]["loss"]
            if aam_margin is None and model_archi["training"]["loss"] == "aam":
                aam_margin = model_archi["training"]["aam_margin"]
            if aam_s is None and model_archi["training"]["loss"] == "aam":
                aam_s = model_archi["training"]["aam_s"]
            if tmp_model_name is None:
                tmp_model_name = model_archi["training"]["tmp_model_name"]
            if best_model_name is None:
                best_model_name = model_archi["training"]["best_model_name"]
            if multi_gpu is None:
                multi_gpu = model_archi["training"]["multi_gpu"]
            if clipping is None:
                clipping = model_archi["training"]["clipping"]

Anthony Larcher's avatar
fix API    
Anthony Larcher committed
925
        if model_name is None:
Anthony Larcher's avatar
Anthony Larcher committed
926
927
928
            model = Xtractor(speaker_number, model_yaml)

         # If we start from an existing model
Anthony Larcher's avatar
Anthony Larcher committed
929
        else:
Anthony Larcher's avatar
Anthony Larcher committed
930
931
932
            # Load the model
            logging.critical(f"*** Load model from = {model_name}")
            checkpoint = torch.load(model_name)
Anthony Larcher's avatar
Anthony Larcher committed
933
            model = Xtractor(speaker_number, model_yaml)
Anthony Larcher's avatar
Anthony Larcher committed
934

Anthony Larcher's avatar
fix API    
Anthony Larcher committed
935
936
            """
            Here we remove all layers that we don't want to reload
Anthony Larcher's avatar
Anthony Larcher committed
937
        
Anthony Larcher's avatar
fix API    
Anthony Larcher committed
938
939
940
941
942
943
944
945
946
947
948
949
950
            """
            pretrained_dict = checkpoint["model_state_dict"]
            for part in reset_parts:
                pretrained_dict = {k: v for k, v in pretrained_dict.items() if not k.startswith(part)}

            new_model_dict = model.state_dict()
            new_model_dict.update(pretrained_dict)
            model.load_state_dict(new_model_dict)

        # Freeze required layers
        for name, param in model.named_parameters():
            if name.split(".")[0] in freeze_parts:
                param.requires_grad = False
Anthony Larcher's avatar
Anthony Larcher committed
951

Anthony Larcher's avatar
Anthony Larcher committed
952
953
954
955
    logging.critical(model)

    logging.critical("model_parameters_count: {:d}".format(
        sum(p.numel()
Anthony Larcher's avatar
debug    
Anthony Larcher committed
956
957
958
959
            for p in model.sequence_network.parameters()
            if p.requires_grad) + \
        sum(p.numel()
            for p in model.before_speaker_embedding.parameters()
Anthony Larcher's avatar
Anthony Larcher committed
960
961
            if p.requires_grad)))

Anthony Larcher's avatar
Anthony Larcher committed
962
    embedding_size = model.embedding_size
Anthony Larcher's avatar
Anthony Larcher committed
963

Anthony Larcher's avatar
Anthony Larcher committed
964
    if torch.cuda.device_count() > 1 and multi_gpu:
Anthony Larcher's avatar
Anthony Larcher committed
965
966
        print("Let's use", torch.cuda.device_count(), "GPUs!")
        model = torch.nn.DataParallel(model)
Anthony Larcher's avatar
Anthony Larcher committed
967
        #model = DDP(model)
Anthony Larcher's avatar
Anthony Larcher committed
968
969
    else:
        print("Train on a single GPU")
Anthony Larcher's avatar
Anthony Larcher committed
970
    model.to(device)
Anthony Larcher's avatar
Anthony Larcher committed
971

Anthony Larcher's avatar
debug    
Anthony Larcher committed
972
973
974
975
    with open(dataset_yaml, "r") as fh:
        dataset_params = yaml.load(fh, Loader=yaml.FullLoader)
        df = pandas.read_csv(dataset_params["dataset_description"])

Anthony Larcher's avatar
Anthony Larcher committed
976
977
978
    if load_batches_from_disk:
        train_batch_fn_format = tmp_batch_dir + "/train/train_{}_batch.h5"
        val_batch_fn_format = tmp_batch_dir + "/val/val_{}_batch.h5"
979

Anthony Larcher's avatar
Anthony Larcher committed
980
981
982
983
984
985
986
    if not load_batches_from_disk or write_batches_to_disk:
        """
        Set the dataloaders according to the dataset_yaml
        
        First we load the dataframe from CSV file in order to split it for training and validation purpose
        Then we provide those two 
        """
Anthony Larcher's avatar
Anthony Larcher committed
987

Anthony Larcher's avatar
minor    
Anthony Larcher committed
988
        if write_batches_to_disk or dataset_params["batch_size"] > 1:
Anthony Larcher's avatar
Anthony Larcher committed
989
990
991
992
            output_format = "numpy"
        else:
            output_format = "pytorch"

Anthony Larcher's avatar
Anthony Larcher committed
993
994
        training_df, validation_df = train_test_split(df, test_size=dataset_params["validation_ratio"])
        torch.manual_seed(dataset_params['seed'])
Anthony Larcher's avatar
Anthony Larcher committed
995
996
997
998
999
1000
        training_set = SpkSet(dataset_yaml,
                              set_type="train",
                              dataset_df=training_df,
                              overlap=dataset_params['train']['overlap'],
                              output_format="pytorch",
                              windowed=True)
For faster browsing, not all history is shown. View entire blame