-
Notifications
You must be signed in to change notification settings - Fork 2
/
data.py
423 lines (348 loc) · 15.6 KB
/
data.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
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
from copy import deepcopy
import json
import os.path as osp
import random
from glob import glob
import numpy as np
import torch
import torch.utils.data as data
import torchvision.transforms as transforms
from transformers import BertTokenizer
from matplotlib.patches import Rectangle, Polygon
from PIL import Image, ImageDraw
from multiprocessing import Manager
class ReferDatasetBert(data.Dataset):
def __init__(self, root, splitset, max_iters=None, transform=None, crop_size=None, label_crop_size=None, scale=True, drop_prob=0):
self.root = root
self.train = 0
self.crop_size = crop_size
self.label_crop_size = label_crop_size
self.transform = transform
self.set = splitset
self.tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
self.vocab = self.tokenizer.vocab
self.drop_prob = drop_prob
self.data_list = glob(osp.join(root, self.set+'_batch','*'))
if not max_iters==None:
self.data_list = self.data_list * int(np.ceil(float(max_iters) / len(self.data_list)))
if max_iters < len(self.data_list):
self.data_list = self.data_list[:max_iters]
self.files = []
def __len__(self):
return len(self.data_list)
def __getitem__(self, index):
sentence, img_id, image, label, im_name = self.get_raw_item(index)
if self.transform is not None:
image = self.transform(image)
target = process_caption_bert(sentence, self.tokenizer, self.drop_prob, self.train)
return image, target, index, img_id
def get_raw_item(self, index):
datafiles = np.load(self.data_list[index])
name = self.data_list[index].split('/')[-1]
name = name.split('.')[0]
image = Image.fromarray(datafiles["im_batch"]).convert('RGB')
label = datafiles["mask_batch"]
sentence = datafiles["sent_batch"][0]
im_name= str(datafiles['im_name_batch'])
img_id = osp.basename(self.data_list[index]).split(".")[0].split("_")[-1]
return sentence, img_id, image, label, im_name
class ReferDatasetBertTexts(data.Dataset):
def __init__(self, root, splitset, max_iters=None, transform=None, crop_size=None,
label_crop_size=None, scale=True, drop_prob=0, num_texts=0):
self.root = root
self.train = 0
self.crop_size = crop_size
self.label_crop_size = label_crop_size
self.transform = transform
self.set = splitset
self.tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
self.vocab = self.tokenizer.vocab
self.drop_prob = drop_prob
self.num_texts = num_texts
with open(osp.join(osp.dirname(root), '%s_imgtotxts.json' % osp.basename(root))) as f:
self.img_txts_dict = json.load(f)
with open(osp.join(osp.dirname(root), '%s_imgtoname.json' % osp.basename(root))) as fp:
self.img_im_name_dict = json.load(fp)
# self.data_list = glob(osp.join(root, self.set+'_batch','*'))
# if not max_iters==None:
# self.data_list = self.data_list * int(np.ceil(float(max_iters) / len(self.data_list)))
# if max_iters < len(self.data_list):
# self.data_list = self.data_list[:max_iters]
self.data_list = [key for key in self.img_im_name_dict.keys()]
if not max_iters==None:
self.data_list = self.data_list * int(np.ceil(float(max_iters) / len(self.data_list)))
if max_iters < len(self.data_list):
self.data_list = self.data_list[:max_iters]
def __len__(self):
return len(self.data_list)
def __getitem__(self, index):
sentences, img_id, image, label, _, num_sent = self.get_raw_item(index)
if self.transform is not None:
image = self.transform(image)
if self.num_texts is not 0:
sentences = random.sample(sentences, self.num_texts) if len(sentences) > self.num_texts else sentences
targets = [process_caption_bert(s, self.tokenizer, self.drop_prob, self.train) for s in sentences]
return image, targets, index, img_id
def get_raw_item(self, index):
img_id = self.img_im_name_dict[self.data_list[index]][0]
datafiles = np.load(osp.join(self.root, self.set+'_batch', img_id))
# name = self.data_list[index].split('/')[-1]
# name = name.split('.')[0]
image = Image.fromarray(datafiles["im_batch"]).convert('RGB')
label = datafiles["mask_batch"]
sentence = datafiles["sent_batch"][0]
img_id = img_id.split(".")[0].split("_")[-1]
im_name = str(datafiles['im_name_batch'])
sentences = self.img_txts_dict[im_name]
num_sent = len(sentences)
return sentences, img_id, image, label, im_name, num_sent
class PhraseDatasetBert(data.Dataset):
def __init__(self, root, splitset, max_iters=None, transform=None, crop_size=None, label_crop_size=None, scale=True, drop_prob=0):
self.root = root
self.train = 0
self.crop_size = crop_size
self.label_crop_size = label_crop_size
self.transform = transform
self.set = splitset
self.tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
# self.vocab = self.tokenizer.vocab
self.drop_prob = drop_prob
self.data_list = []
with open(osp.join("./data/phrasecut/", "refer_%s_ris.json") % splitset, "r") as json_file:
ref_tasks_dict = json.load(json_file)
for i, (key, values) in enumerate(ref_tasks_dict.items()):
if splitset == 'train':
self.data_list.append((key, values['phrase'], None))
elif splitset == 'val':
image = Image.open(osp.join(root,'images', key.split('__')[0]+'.jpg')).convert('RGB')
polygons = []
for ps in values['Polygons']:
polygons += ps
label = polygons_to_mask(polygons, image.size[0], image.size[1])
self.data_list.append((key, values['phrase'], label))
# del label, image, polygons
if not max_iters == None:
if i > max_iters-1:
break
del ref_tasks_dict, json_file
self.data_ids = [i for i in range(len(self.data_list))]
manager = Manager()
self.data_list = manager.list(self.data_list)
# self.data_ids = manager.list(self.data_ids)
def __len__(self):
return len(self.data_list)
def __getitem__(self, index):
sentence, img_id, image, label, _ = self.get_raw_item(index)
if self.transform is not None:
image = self.transform(image)
target = process_caption_bert(sentence, self.tokenizer, self.drop_prob, self.train)
del label, sentence
return image, target, index, img_id
def get_raw_item(self, index):
ref_id, sentence, label = self.data_list[self.data_ids[index]]
image = Image.open(osp.join(self.root,'images', ref_id.split('__')[0]+'.jpg')).convert('RGB')
return sentence, ref_id, image, label, None
def polygons_to_mask(polygons, w, h):
p_mask = np.zeros((h, w))
for polygon in polygons:
if len(polygon) < 2:
continue
p = []
for x, y in polygon:
p.append((int(x), int(y)))
# img = Image.new('L', (w, h), 0)
with Image.new('L', (w, h), 0) as im:
ImageDraw.Draw(im).polygon(p, outline=1, fill=1)
mask = np.array(im)
p_mask += mask
del im, mask, p
p_mask = p_mask > 0
return p_mask
def process_caption_bert(caption, tokenizer, drop_prob, train):
output_tokens = []
deleted_idx = []
tokens = tokenizer.basic_tokenizer.tokenize(caption)
for i, token in enumerate(tokens):
sub_tokens = tokenizer.wordpiece_tokenizer.tokenize(token)
prob = random.random()
if prob < drop_prob and train: # mask/remove the tokens only during training
prob /= drop_prob
# 50% randomly change token to mask token
if prob < 0.5:
for sub_token in sub_tokens:
output_tokens.append("[MASK]")
# 10% randomly change token to random token
elif prob < 0.6:
for sub_token in sub_tokens:
output_tokens.append(random.choice(list(tokenizer.vocab.keys())))
# -> rest 10% randomly keep current token
else:
for sub_token in sub_tokens:
output_tokens.append(sub_token)
deleted_idx.append(len(output_tokens) - 1)
else:
for sub_token in sub_tokens:
# no masking token (will be ignored by loss function later)
output_tokens.append(sub_token)
if len(deleted_idx) != 0:
output_tokens = [output_tokens[i] for i in range(len(output_tokens)) if i not in deleted_idx]
output_tokens = ['[CLS]'] + output_tokens + ['[SEP]']
target = tokenizer.convert_tokens_to_ids(output_tokens)
target = torch.Tensor(target)
return target
def collate_fn(data):
"""Build mini-batch tensors from a list of (image, sentence) tuples.
Args:
data: list of (image, sentence) tuple.
- image: torch tensor of shape (3, 256, 256) or (?, 3, 256, 256).
- sentence: torch tensor of shape (?); variable length.
Returns:
images: torch tensor of shape (batch_size, 3, 256, 256) or
(batch_size, padded_length, 3, 256, 256).
targets: torch tensor of shape (batch_size, padded_length).
lengths: list; valid length for each padded sentence.
"""
# Sort a data list by sentence length
data.sort(key=lambda x: len(x[1]), reverse=True)
images, sentences, ids, img_ids = zip(*data)
# Merge images (convert tuple of 3D tensor to 4D tensor)
images = torch.stack(images, 0)
# Merge sentences (convert tuple of 1D tensor to 2D tensor)
cap_lengths = torch.tensor([len(cap) for cap in sentences])
targets = torch.zeros(len(sentences), max(cap_lengths)).long()
for i, cap in enumerate(sentences):
end = cap_lengths[i]
targets[i, :end] = cap[:end]
return images, targets, cap_lengths, ids
def collate_fn_fast(data):
"""
input : List of tuples. Each tuple is a output of __getitem__ of the dataset
output : Collated tensor
"""
# Sort a data list by sentence length
images, sentences, img_ids, _ = zip(*data)
# image, sentences, index, img_id
# compute the number of captions in each images and create match label from it
flatten_sentences = [sentence for img in list(sentences) for sentence in img]
flatten_sentences_len = [len(sentence) for sentence in flatten_sentences]
n_sents_for_img = [len(sents) for sents in list(sentences)] #len = batch
org_len, org_sen = flatten_sentences_len, flatten_sentences
caption_data = list(zip(flatten_sentences_len, flatten_sentences))
sorted_idx = sorted(range(len(caption_data)), key=lambda x: caption_data[x][0], reverse=True)
recovery_idx = sorted(range(len(caption_data)), key=lambda x: sorted_idx[x], reverse=False)
caption_data.sort(key=lambda x: x[0], reverse=True)
flatten_sentences_len, flatten_sentences = zip(*caption_data)
# Merge images (convert tuple of 3D tensor to 4D tensor)
images = torch.stack(images, 0)
# Merge sentences (convert tuple of 1D tensor to 2D tensor)
sentences_len = torch.tensor(flatten_sentences_len)
recovery_idx = torch.tensor(recovery_idx)
n_sents_for_img = torch.tensor(n_sents_for_img)
padded_sentences = torch.zeros(len(flatten_sentences), max(sentences_len)).long()
for i, cap in enumerate(flatten_sentences):
end = sentences_len[i]
padded_sentences[i, :end] = cap[:end]
return images, padded_sentences, sentences_len, recovery_idx, n_sents_for_img, img_ids
def get_loader_single(
data_name, split, root, transform, fast_batch,
batch_size=128, shuffle=True, num_workers=2, max_iters=None, num_texts=0):
"""Returns torch.utils.data.DataLoader for custom coco dataset."""
if 'coco' == data_name:
if fast_batch:
dataset = ReferDatasetBertTexts(
root=root,
splitset = split,
transform=transform,
num_texts=num_texts)
else:
dataset = ReferDatasetBert(
root=root,
splitset = split,
transform=transform)
elif 'phrasecut' == data_name:
dataset = PhraseDatasetBert(
root=root,
splitset = split,
transform=transform,
max_iters=max_iters
)
else:
assert NotImplementedError
collate = collate_fn_fast if fast_batch else collate_fn
# Data loader
data_loader = torch.utils.data.DataLoader(
dataset=dataset,
batch_size=batch_size,
shuffle=shuffle,
pin_memory=True,
num_workers=num_workers,
# prefetch_factor = 2,
persistent_workers=True,
collate_fn=collate)
return data_loader
def get_image_transform(split_name, img_backbone, crop_size, use_aug):
if 'vit' in img_backbone:
normalizer = transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5])
elif 'res' in img_backbone:
normalizer = transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
else:
raise NotImplementedError
t_list = []
if split_name == 'train':
if use_aug:
t_list = [
transforms.RandomResizedCrop(size=(crop_size, crop_size), scale=(0.8, 1.0), ratio=(0.8, 1.2)),
# transforms.RandomHorizontalFlip()
]
else:
t_list = [transforms.Resize(size=(crop_size, crop_size))]
elif split_name == 'val':
t_list = [
transforms.Resize(size=(crop_size, crop_size))
]
t_end = [
transforms.ToTensor(),
normalizer
]
transform = transforms.Compose(t_list + t_end)
return transform
def get_train_loader(args):
transform = get_image_transform('train', args.img_backbone, args.crop_size, args.use_aug)
return get_loader_single(
data_name=args.data_name,
split=args.data_split,
root=args.data_path,
transform=transform,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.workers,
fast_batch=args.fast_batch,
num_texts=args.num_texts,
)
def get_test_loader(args):
transform = get_image_transform('val', args.img_backbone, args.crop_size, args.use_aug)
return get_loader_single(
data_name=args.data_name,
split='val',
root=args.data_path,
transform=transform,
batch_size=args.batch_size,
shuffle=False,
num_workers=1,
max_iters = 10000,
fast_batch=False,
num_texts=args.num_texts,
)
def get_train_pseudo_loader(args):
transform = get_image_transform('train', args.img_backbone, args.crop_size, use_aug=False)
return get_loader_single(
data_name=args.data_name,
split=args.data_split,
root=args.data_path,
transform=transform,
batch_size=args.batch_size,
shuffle=False,
num_workers=1,
fast_batch=False,
num_texts=args.num_texts,
)