-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathoverwrite_rnn.py
131 lines (113 loc) · 5.91 KB
/
overwrite_rnn.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
import argparse
import os
import pickle
import pandas as pd
import torch
import torch.nn as nn
from data.util import get_trec_dataset, get_mnist_dataset
from models.layers.util import sign_to_str, calculate_ber
from models.util import seed_everything
from trainer.rnn import PrivateTrainer
from trainer.util import sequence_to_text
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--output-dir', type=str, dest='output_dir', help='output folder')
parser.add_argument('--seed', type=int, dest='seed', default=1234, help='seed for experiment')
parser.add_argument('--epochs', type=int, dest='epochs', default=5, help='number of epochs')
parser.add_argument('--batch-size', type=int, dest='batch_size', default=64, help='batch size per steps')
parser.add_argument('--max-sentence-length', type=int, dest='max_sentence_length', default=30,
help='max sentence length, only used in nlp task')
parser.add_argument('--pretrained-path', type=str, dest='pretrained_path', help='path to saved pretrained model',
required=True)
args = parser.parse_args()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
seed = args.seed
seed_everything(seed)
batch_size = args.batch_size
epochs = args.epochs
max_sentence_length = args.max_sentence_length
lr = 0.00001
save_dir = args.output_dir
os.makedirs(save_dir, exist_ok=True)
with open(os.path.join(args.pretrained_path, 'keyed_kwargs_{}.pickle'.format(seed)), 'rb') as f:
keyed_kwargs = pickle.load(f)
dataset = keyed_kwargs['dataset']
trigger_size = keyed_kwargs['trigger_size']
trigger_batch_size = keyed_kwargs['trigger_batch_size']
vocab = None
pad_idx = 0
if os.path.isfile(os.path.join(args.pretrained_path, 'trigger_dataloader_{}.pth'.format(seed))):
print('found trigger dataset')
trigger_dataloader = torch.load(os.path.join(args.pretrained_path, 'trigger_dataloader_{}.pth'.format(seed)))
if dataset == 'trec':
train_dataloader, valid_dataloader, _, vocab = get_trec_dataset(num_workers=2,
batch_size=batch_size,
trigger_size=trigger_size,
trigger_batch_size=trigger_batch_size,
max_sentence_length=max_sentence_length)
else:
train_dataloader, valid_dataloader, _ = get_mnist_dataset(num_workers=2,
batch_size=batch_size,
trigger_size=trigger_size,
trigger_batch_size=trigger_batch_size)
else:
trigger_dataloader = None
if dataset == 'trec':
train_dataloader, valid_dataloader, vocab = get_trec_dataset(num_workers=2, batch_size=batch_size,
max_sentence_length=max_sentence_length)
else:
train_dataloader, valid_dataloader = get_mnist_dataset(num_workers=2, batch_size=batch_size)
if dataset == 'trec':
num_words = len(vocab.itos)
pad_idx = vocab.stoi['<pad>']
model = torch.load(os.path.join(args.pretrained_path, 'model_{}.pth'.format(seed)))
# remove sign loss
model.rnn.sign_loss = None
old_key = model.key.cpu().clone()
# overwrite with new key
if 'overwrite_key' in keyed_kwargs:
keyed_kwargs['key'] = keyed_kwargs.pop('overwrite_key')
else:
# prevent repeating key due to same seed
if dataset == 'trec':
keyed_kwargs['key'] = torch.randint(max(0, num_words - 2000), num_words, (8, max_sentence_length))
keyed_kwargs['key'] = torch.randint(max(0, num_words - 2000), num_words, (8, max_sentence_length))
print('new key:')
print(sequence_to_text(keyed_kwargs['key'], vocab))
else:
keyed_kwargs['key'] = torch.randn((8, 28, 28))
keyed_kwargs['key'] = torch.randn((8, 28, 28))
model.set_key(keyed_kwargs['key'].to(device))
optimizer = torch.optim.Adam(model.parameters(), lr)
criterion = nn.CrossEntropyLoss()
trainer = PrivateTrainer(model, optimizer, criterion, device)
train_res, test_res, trigger_res = [], [], []
for e in range(1, epochs + 1):
tra = trainer.train(e, train_dataloader)
tes = trainer.test(valid_dataloader)
if trigger_dataloader:
tri = trainer.test(trigger_dataloader, msg='Trigger testing')
trigger_res.append(tri)
train_res.append(tra)
test_res.append(tes)
train_df = pd.DataFrame(train_res)
test_df = pd.DataFrame(test_res)
train_df.to_csv(os.path.join(save_dir, 'train_{}.csv'.format(seed)))
test_df.to_csv(os.path.join(save_dir, 'valid_{}.csv'.format(seed)))
print('average training time per epoch:', train_df['time'].mean())
if trigger_dataloader:
trigger_df = pd.DataFrame(trigger_res)
trigger_df.to_csv(os.path.join(save_dir, 'trigger_{}.csv'.format(seed)))
torch.save(model, os.path.join(save_dir, 'model_{}.pth'.format(seed)))
print('*' * 50)
print('restoring old keys:')
model.set_key(old_key.to(device))
trainer = PrivateTrainer(model, optimizer, criterion, device)
te = trainer.test(valid_dataloader)
tri = trainer.test(trigger_dataloader, msg='Trigger testing')
print('\nsignature:')
try:
print(sign_to_str(model.get_signature().cpu().detach().numpy(), len(keyed_kwargs['signature'])))
except UnicodeError:
pass
print('ber:', calculate_ber(model.get_signature().cpu().detach().numpy(), keyed_kwargs['signature']))