-
Notifications
You must be signed in to change notification settings - Fork 14
/
Copy pathdata_utils.py
180 lines (156 loc) · 7.33 KB
/
data_utils.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
import random
import numpy as np
import torch
import torch.utils.data
import os
import layers
from utils import load_wav_to_torch, load_filepaths_and_text
from text import text_to_sequence
class TextMelLoader(torch.utils.data.Dataset):
"""
1) loads audio,text pairs
2) normalizes text and converts them to sequences of one-hot vectors
3) computes mel-spectrograms from audio files.
"""
def __init__(self, audiopaths, hparams):
self.inputs = audiopaths[0]
self.outputs = audiopaths[1]
self.text_cleaners = hparams.text_cleaners
self.max_wav_value = hparams.max_wav_value
self.sampling_rate = hparams.sampling_rate
self.load_mel_from_disk = hparams.load_mel_from_disk
self.stft = layers.TacotronSTFT(
hparams.filter_length, hparams.hop_length, hparams.win_length,
hparams.n_mel_channels, hparams.sampling_rate, hparams.mel_fmin,
hparams.mel_fmax)
# random.seed(hparams.seed)
# random.shuffle(self.audiopaths_and_text)
def get_mel_spec_pair(self, index):
# separate filename and text
# lin = self.get_spec(self.outputs[index])
# mel = self.get_mel(self.inputs[index])
inputs = self.get_mel(self.inputs[index])
outputs = self.get_mel(self.outputs[index])
return (inputs,outputs)
def get_mel(self, filename):
if not self.load_mel_from_disk:
audio, sampling_rate = load_wav_to_torch(filename)
# if sampling_rate != self.stft.sampling_rate:
# raise ValueError("{} {} SR doesn't match target {} SR".format(
# sampling_rate, self.stft.sampling_rate))
audio_norm = audio / self.max_wav_value
audio_norm = audio_norm.unsqueeze(0)
audio_norm = torch.autograd.Variable(audio_norm, requires_grad=False)
melspec = self.stft.mel_spectrogram(audio_norm)
melspec = torch.squeeze(melspec, 0)
else:
melspec = torch.from_numpy(np.load(filename))
assert melspec.size(0) == self.stft.n_mel_channels, (
'Mel dimension mismatch: given {}, expected {}'.format(
melspec.size(0), self.stft.n_mel_channels))
return melspec
def get_spec(self, filename):
if not self.load_mel_from_disk:
audio, sampling_rate = load_wav_to_torch(filename)
if sampling_rate != self.stft.sampling_rate:
raise ValueError("{} {} SR doesn't match target {} SR".format(
sampling_rate, self.stft.sampling_rate))
audio_norm = audio / self.max_wav_value
audio_norm = audio_norm.unsqueeze(0)
audio_norm = torch.autograd.Variable(audio_norm, requires_grad=False)
spec = self.stft.spectrogram(audio_norm)
spec = torch.squeeze(spec, 0)
else:
spec = torch.from_numpy(np.load(filename))
# assert melspec.size(0) == self.stft.n_mel_channels, (
# 'Mel dimension mismatch: given {}, expected {}'.format(
# melspec.size(0), self.stft.n_mel_channels))
return spec
# def get_text(self, text):
# text_norm = torch.IntTensor(text_to_sequence(text, self.text_cleaners))
# return text_norm
def __getitem__(self, index):
return self.get_mel_spec_pair(index)
def __len__(self):
return len(self.inputs)
class TextMelCollate():
""" Zero-pads model inputs and targets based on number of frames per setep
"""
def __init__(self, n_frames_per_step):
self.n_frames_per_step = n_frames_per_step
def __call__(self, batch):
"""Collate's training batch from normalized text and mel-spectrogram
PARAMS
------
batch: [text_normalized, mel_normalized]
"""
# Right zero-pad mel-spec
# num_mels = batch[0][1].size(0)
# max_input_len = max([x[0].size(1) for x in batch])
# if max_input_len % self.n_frames_per_step != 0:
# max_input_len += self.n_frames_per_step - max_input_len % self.n_frames_per_step
# assert max_input_len % self.n_frames_per_step == 0
# # include mel padded and gate padded
# mel_padded = torch.FloatTensor(len(batch), num_mels, max_input_len)
# mel_padded.zero_()
# gate_padded = torch.FloatTensor(len(batch), max_input_len)
# gate_padded.zero_()
# input_lengths = torch.LongTensor(len(batch))
# for i in range(len(batch)):
# mel = batch[i][0]
# mel_padded[i, :, :mel.size(1)] = mel
# gate_padded[i, mel.size(1)-1:] = 1
# input_lengths[i] = mel.size(1)
# # input_lengths, ids_sorted_decreasing = torch.sort(
# # torch.LongTensor([len(x[0]) for x in batch]),
# # dim=0, descending=True)
# num_dims = batch[0][1].size(0)
# max_target_len = max([x[1].size(1) for x in batch])
# spec_padded = torch.FloatTensor(len(batch), num_dims, max_target_len)
# spec_padded.zero_()
# # gate_padded = torch.FloatTensor(len(batch), max_target_len)
# # gate_padded.zero_()
# output_lengths = torch.LongTensor(len(batch))
# for i in range(len(batch)):
# spec = batch[i][1]
# spec_padded[i, :, :spec.size(1)] = spec
# # gate_padded[i, mel.size(1)-1:] = 1
# output_lengths[i] = spec.size(1)
# return mel_padded, gate_padded,input_lengths, spec_padded, \
# output_lengths
num_mels = batch[0][0].size(0)
# max_input_len = max([x[0].size(1) for x in batch])
input_lengths, ids_sorted_decreasing = torch.sort(
torch.LongTensor([x[0].size(1) for x in batch]),
dim=0, descending=True)
max_input_len = input_lengths[0]
if max_input_len % self.n_frames_per_step != 0:
max_input_len += self.n_frames_per_step - max_input_len % self.n_frames_per_step
assert max_input_len % self.n_frames_per_step == 0
# include mel padded and gate padded
input_padded = torch.FloatTensor(len(batch), num_mels, max_input_len)
input_padded.zero_()
# gate_padded = torch.FloatTensor(len(batch), max_target_len)
# gate_padded.zero_()
for i in ids_sorted_decreasing:
mel = batch[i][0]
input_padded[i, :, :mel.size(1)] = mel
# gate_padded[i, mel.size(1)-1:] = 1
num_mels = batch[0][1].size(0)
max_target_len = max([x[1].size(1) for x in batch])
if max_target_len % self.n_frames_per_step != 0:
max_target_len += self.n_frames_per_step - max_target_len % self.n_frames_per_step
assert max_target_len % self.n_frames_per_step == 0
# include mel padded and gate padded
mel_padded = torch.FloatTensor(len(batch), num_mels, max_target_len)
mel_padded.zero_()
gate_padded = torch.FloatTensor(len(batch), max_target_len)
gate_padded.zero_()
output_lengths = torch.LongTensor(len(batch))
for i in ids_sorted_decreasing:
mel = batch[i][1]
mel_padded[i, :, :mel.size(1)] = mel
gate_padded[i, mel.size(1)-1:] = 1
output_lengths[i] = mel.size(1)
return input_padded, input_lengths, mel_padded, gate_padded, \
output_lengths