-
Notifications
You must be signed in to change notification settings - Fork 2
/
data_loaders.py
213 lines (171 loc) · 8.14 KB
/
data_loaders.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
import os
import numpy as np
import torch
import torchvision
import torchvision.transforms as transforms
from datasets import load_dataset
from transformers.data.data_collator import DataCollatorWithPadding
from models.dlrm.dlrm_data_pytorch import CriteoDataset, collate_wrapper_criteo
import torch
from torchvision import datasets
class ImageFolderWithPaths(datasets.ImageFolder):
"""Custom dataset that includes image file paths. Extends
torchvision.datasets.ImageFolder
"""
def __getitem__(self, index):
original_tuple = super(ImageFolderWithPaths, self).__getitem__(index)
path = self.imgs[index][0]
tuple_with_path = (original_tuple + (path,))
return tuple_with_path
def get_dlrm_dataloaders(data_path, train_batch_size, val_batch_size):
raw_file = os.path.join(data_path, 'train.txt')
processed_file = os.path.join(data_path, 'kaggleAdDisplayChallenge_processed.npz')
train_data = CriteoDataset('kaggle', -1, 0.875, 'total', "train", raw_file, processed_file, True)
train_loader = torch.utils.data.DataLoader(
train_data,
batch_size=train_batch_size,
shuffle=False,
collate_fn=collate_wrapper_criteo,
pin_memory=False,
drop_last=False, # True
)
test_data = CriteoDataset('kaggle', -1, 0.875, 'total', "test", raw_file, processed_file, True)
test_loader = torch.utils.data.DataLoader(
test_data,
batch_size=val_batch_size,
shuffle=False,
collate_fn=collate_wrapper_criteo,
pin_memory=False,
drop_last=False, # True
)
return train_loader, test_loader
def get_imagenet_dataloaders(data_path, train_batch_size, val_batch_size, train_data_path, val_data_path, args, train_subset_size=None):
normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
if train_data_path is None:
traindir = os.path.join(data_path, 'imagenet/train_subset')
else:
traindir = train_data_path
print("Training on {}".format(traindir))
crop_size = 299 if args.model == "inceptionv3" else 224
dataset_train = ImageFolderWithPaths(
traindir,
transforms.Compose([
transforms.RandomResizedCrop(crop_size),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
normalize,
]))
num_epochs = -1
if train_subset_size is not None and train_subset_size != len(dataset_train):
full_batches = np.ceil(len(dataset_train)/train_batch_size)
full_batches = full_batches if args.train_batches == -1 else min(full_batches, args.train_batches)
subset_batches = np.ceil(train_subset_size/train_batch_size)
subset_batches = subset_batches if args.train_batches == -1 else min(subset_batches, args.train_batches)
num_epochs = int(np.ceil(full_batches/subset_batches))
np.random.seed(1234)
train_subset_indices = np.random.choice(len(dataset_train), train_subset_size)
dataset_train = torch.utils.data.Subset(dataset_train, train_subset_indices)
print("Training with {} samples".format(len(dataset_train)))
data_loader_train = torch.utils.data.DataLoader(
dataset_train, batch_size=train_batch_size, num_workers=4,
shuffle=True, pin_memory=True, drop_last=False)
if val_data_path is None:
valdir = os.path.join(data_path, 'full_imagenet/val_2012/val_formatted')
else:
valdir = val_data_path
print("Evaluating on {}".format(valdir))
resize_size = 299 if args.model == "inceptionv3" else 256
dataset_val = ImageFolderWithPaths(
valdir,
transforms.Compose([
transforms.Resize(resize_size),
transforms.CenterCrop(crop_size),
transforms.ToTensor(),
normalize,
]))
data_loader_val = torch.utils.data.DataLoader(
dataset_val, batch_size=val_batch_size, num_workers=4,
shuffle=False, pin_memory=True, drop_last=False)
return data_loader_train, data_loader_val, num_epochs
def get_glue_dataloaders(tokenizer, train_batch_size, val_batch_size):
datasets = load_dataset("glue", "mnli")
sentence1_key, sentence2_key = ("premise", "hypothesis")
padding = "max_length"
max_length = 128
label_to_id = None
def preprocess_function(examples):
# Tokenize the texts
args = (
(examples[sentence1_key],) if sentence2_key is None else (examples[sentence1_key], examples[sentence2_key])
)
result = tokenizer(*args, padding=padding, max_length=max_length, truncation=True)
# Map labels to IDs (not necessary for GLUE tasks)
if label_to_id is not None and "label" in examples:
result["label"] = [label_to_id[l] for l in examples["label"]]
return result
datasets = datasets.map(preprocess_function, batched=True)
train_dataset = datasets["train"]
val_matched_dataset = datasets["validation_matched"]
val_mismatched_dataset = datasets["validation_mismatched"]
train_dataset.set_format(type=train_dataset.format["type"], columns=['attention_mask', 'input_ids', 'label', 'token_type_ids'])
val_matched_dataset.set_format(type=val_matched_dataset.format["type"], columns=['attention_mask', 'input_ids', 'label', 'token_type_ids'])
val_mismatched_dataset.set_format(type=val_matched_dataset.format["type"], columns=['attention_mask', 'input_ids', 'label', 'token_type_ids'])
train_loader = torch.utils.data.DataLoader(
train_dataset,
batch_size=train_batch_size,
collate_fn = DataCollatorWithPadding(tokenizer),
shuffle=True,
pin_memory=False,
drop_last=False,
)
test_loader_1 = torch.utils.data.DataLoader(
val_matched_dataset,
batch_size=val_batch_size,
collate_fn=DataCollatorWithPadding(tokenizer),
shuffle=False,
pin_memory=False,
drop_last=False,
)
test_loader_2 = torch.utils.data.DataLoader(
val_mismatched_dataset,
batch_size=val_batch_size,
collate_fn=DataCollatorWithPadding(tokenizer),
shuffle=False,
pin_memory=False,
drop_last=False, # True
)
return train_loader, (test_loader_1, test_loader_2)
def prepare_data_loaders(dataset, data_path, train_batch_size, val_batch_size,
train_data_path, val_data_path, args, train_subset_size=None, tokenizer=None):
if dataset == 'criteo':
train_loader, test_loader = get_dlrm_dataloaders(data_path, train_batch_size, val_batch_size)
num_epochs = args.epochs
elif dataset == 'cifar10':
train_loader, test_loader = get_cifar10_dataloaders(data_path, train_batch_size, val_batch_size)
num_epochs = args.epochs
elif dataset == 'mnli':
train_loader, test_loader = get_glue_dataloaders(tokenizer, train_batch_size, val_batch_size)
num_epochs = args.epochs
else:
train_loader, test_loader, num_epochs = get_imagenet_dataloaders(data_path, train_batch_size,
val_batch_size, train_data_path, val_data_path, args, train_subset_size)
return train_loader, test_loader, num_epochs
def get_cifar10_dataloaders(data_path, train_batch_size, val_batch_size):
mean = [0.4914, 0.4822, 0.4465]
std = [0.2023, 0.1994, 0.2010]
dataset_train = torchvision.datasets.CIFAR10(root="data", train=True,
transform=transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(), transforms.ToTensor(),
transforms.Normalize(mean, std)]), download=True)
data_loader_train = torch.utils.data.DataLoader(
dataset_train, batch_size=train_batch_size, num_workers=4,
shuffle=True, pin_memory=True, drop_last=False)
dataset_val = torchvision.datasets.CIFAR10(root="data", train=False,
transform=([transforms.ToTensor(),
transforms.Normalize(mean, std)]), download=True)
data_loader_val = torch.utils.data.DataLoader(
dataset_val, batch_size=val_batch_size, num_workers=4,
shuffle=False, pin_memory=True, drop_last=False)
return data_loader_train, data_loader_val