Spaces:
Build error
Build error
File size: 9,995 Bytes
62e9ca6 |
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 |
# ----------------------------------------------------------------------------
# VatLM: Visual-Audio-Text Pre-Training with Unified Masked Prediction for Speech Representation Learning
# Github source: https://github.com/microsoft/SpeechT5/tree/main/VATLM
# Code based on fairseq: https://github.com/facebookresearch/fairseq and av_hubert: https://github.com/facebookresearch/av_hubert
#
# Copyright (c) 2022 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# ----------------------------------------------------------------------------
import itertools
import logging
import os
import sys
import time
from typing import Any, List, Optional, Union
import numpy as np
import torch
import torch.nn.functional as F
from fairseq.data import data_utils
from fairseq.data.fairseq_dataset import FairseqDataset
DBG=True if len(sys.argv) == 1 else False
if DBG:
import utils as custom_utils
logging.basicConfig(
format="%(asctime)s | %(levelname)s | %(name)s | %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
level=os.environ.get("LOGLEVEL", "DEBUG").upper(),
stream=sys.stdout,
)
else:
from . import utils as custom_utils
logger = logging.getLogger(__name__)
def load_text(manifest_path, max_keep, min_keep, frame_rate, label_paths, label_rates, tol=0.1):
n_long, n_short, n_unaligned = 0, 0, 0
names, inds, sizes = [], [], []
dur_from_label_list = []
with open(manifest_path) as f:
for ind, line in enumerate(f):
items = line.strip().split("\t")
frames = items[0]
sz = int(frames)
if min_keep is not None and sz < min_keep:
n_short += 1
elif max_keep is not None and sz > max_keep:
n_long += 1
else:
inds.append(ind)
sizes.append(sz)
logger.info(
(
f"max_keep={max_keep}, min_keep={min_keep}, "
f"loaded {len(inds)}, skipped {n_short} short and {n_long} long"
f"longest-loaded={max(sizes)}, shortest-loaded={min(sizes)}"
)
)
return inds, sizes
def load_label(label_path, inds):
with open(label_path) as f:
labels = [line.rstrip() for line in f]
labels = [labels[i] for i in inds]
return labels
def load_phone_label(tsv, inds):
with open(tsv) as f:
labels = [line.rstrip() for line in f.readlines()]
labels = [labels[i] for i in inds]
return labels
def load_label_offset(label_path, inds):
with open(label_path) as f:
code_lengths = [len(line.encode("utf-8")) for line in f]
offsets = list(itertools.accumulate([0] + code_lengths))
offsets = [(offsets[i], offsets[i + 1]) for i in inds]
return offsets
class TextHubertDataset(FairseqDataset):
def __init__(
self,
manifest_path: str,
sample_rate: float,
label_paths: List[str],
label_rates: Union[List[float], float], # -1 for sequence labels
pad_list: List[str],
eos_list: List[str],
label_processors: Optional[List[Any]] = None,
phone_sequence_processors: Optional[List[Any]] = None,
max_keep_sample_size: Optional[int] = None,
min_keep_sample_size: Optional[int] = None,
max_sample_size: Optional[int] = None,
shuffle: bool = True,
pad_audio: bool = False,
normalize: bool = False,
store_labels: bool = True,
single_target: bool = False,
stack_order_audio: int=1,
skip_verify: bool=False,
is_s2s=False,
):
self.label_rates = (
[label_rates for _ in range(len(label_paths))]
if isinstance(label_rates, int)
else label_rates
)
inds, self.sizes = load_text(manifest_path, max_keep_sample_size, min_keep_sample_size, frame_rate=sample_rate, label_paths=label_paths, label_rates=self.label_rates)
self.sample_rate = sample_rate
self.stack_order_audio = stack_order_audio
self.shuffle = shuffle
self.num_labels = len(label_paths)
self.pad_list = pad_list
self.eos_list = eos_list
self.label_processors = label_processors
self.phone_processors = phone_sequence_processors
self.single_target = single_target
self.store_labels = store_labels
self.is_s2s = is_s2s
if store_labels:
self.label_list = [load_label(p, inds) for p in label_paths]
self.phone_list = [load_phone_label(p, inds) for p in [manifest_path]]
else:
self.label_paths = label_paths
self.label_offsets_list = [
load_label_offset(p, inds) for p in label_paths
]
self.max_sample_size = (
max_sample_size if max_sample_size is not None else sys.maxsize
)
self.pad_audio = pad_audio
self.normalize = normalize
def get_label(self, index, label_idx):
if self.store_labels:
label = self.label_list[label_idx][index]
else:
with open(self.label_paths[label_idx]) as f:
offset_s, offset_e = self.label_offsets_list[label_idx][index]
f.seek(offset_s)
label = f.read(offset_e - offset_s)
if self.label_processors is not None:
label = self.label_processors[label_idx](label)
return label
def get_labels(self, index):
return [self.get_label(index, i) for i in range(self.num_labels)]
def get_phone(self, index, label_idx):
label = self.phone_list[label_idx][index]
if self.phone_processors is not None:
label = self.phone_processors[label_idx](label)
return label
def get_phones(self, index):
return [self.get_phone(index, i) for i in range(1)]
def __getitem__(self, index):
labels = self.get_labels(index)
phone_sequence_list = self.get_phones(index)
return {"id": index, "label_list": labels, "phone_sequence_list": phone_sequence_list}
def __len__(self):
return len(self.sizes)
def collater(self, samples):
samples = [s for s in samples if s["id"] is not None]
if len(samples) == 0:
return {}
targets_by_label = [
[s["label_list"][i] for s in samples]
for i in range(self.num_labels)
]
targets_list, lengths_list, ntokens_list = self.collater_label(
targets_by_label,
)
phone_sequence_list = [s["phone_sequence_list"] for s in samples]
if phone_sequence_list[0] is None:
phone_sequence_list = None
targets_by_phone_label = [
[s["phone_sequence_list"][i] for s in samples]
for i in range(self.num_labels)
]
targets_phone_list, lengths_phone_list, ntokens_phone_list = self.collater_phone_label(
targets_by_phone_label,
)
net_input = {"source": None}
batch = {
"id": torch.LongTensor([s["id"] for s in samples]),
"net_input": net_input,
}
if self.single_target:
batch["target_lengths"] = lengths_list[0]
batch["ntokens"] = ntokens_list[0]
if self.is_s2s:
batch['target'], net_input['prev_output_tokens'] = targets_list[0][0], targets_list[0][1]
else:
batch["target"] = targets_list[0]
else:
batch["target_lengths_list"] = lengths_list
batch["ntokens_list"] = ntokens_list
batch["target_list"] = targets_list
batch["extra_text_phone_list"] = targets_phone_list
return batch
def collater_frm_label(
self, targets, label_rate, pad
):
lengths = torch.LongTensor([len(t) for t in targets])
ntokens = lengths.sum().item()
targets = data_utils.collate_tokens(
targets, pad_idx=pad, left_pad=False
)
return targets, lengths, ntokens
def collater_frm_phone_label(
self, targets, pad
):
lengths = torch.LongTensor([len(t) for t in targets])
ntokens = lengths.sum().item()
targets = data_utils.collate_tokens(
targets, pad_idx=pad, left_pad=False
)
return targets, lengths, ntokens
def collater_label(self, targets_by_label,):
targets_list, lengths_list, ntokens_list = [], [], []
itr = zip(targets_by_label, self.label_rates, self.pad_list)
for targets, label_rate, pad in itr:
targets, lengths, ntokens = self.collater_frm_label(
targets, label_rate, pad
)
targets_list.append(targets)
lengths_list.append(lengths)
ntokens_list.append(ntokens)
return targets_list, lengths_list, ntokens_list
def collater_phone_label(self, targets_by_label):
targets_list, lengths_list, ntokens_list = [], [], []
itr = zip(targets_by_label, self.label_rates, self.pad_list)
for targets, label_rate, pad in itr:
targets, lengths, ntokens = self.collater_frm_phone_label(
targets, pad
)
targets_list.append(targets)
lengths_list.append(lengths)
ntokens_list.append(ntokens)
return targets_list, lengths_list, ntokens_list
def num_tokens(self, index):
return self.size(index)
def size(self, index):
if self.pad_audio:
return self.sizes[index]
return min(self.sizes[index], self.max_sample_size)
def ordered_indices(self):
if self.shuffle:
order = [np.random.permutation(len(self))]
else:
order = [np.arange(len(self))]
order.append(self.sizes)
return np.lexsort(order)[::-1]
|