-
Notifications
You must be signed in to change notification settings - Fork 30
/
main_biased_mnist.py
138 lines (123 loc) · 5.59 KB
/
main_biased_mnist.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
"""ReBias
Copyright (c) 2020-present NAVER Corp.
MIT license
Entry point of Biased-MNIST experiments.
This script provides full implementations including
- Various methods (ReBias, Vanilla, Biased, LearnedMixIn, RUBi)
- Target network: Stacked convolutional networks (kernel_size=7)
- Biased network: Stacked convolutional networks (kernel_size=1)
- We do not provide HEX implementation here. See README.md for details.
- Controllable Biased-MNIST experiments by --train_correlation option.
- Please see datasets/colour_mnist.py for details.
Usage:
python main_biased_mnist.py --root /path/to/your/dataset --train_correlation 0.999
"""
import fire
from datasets.colour_mnist import get_biased_mnist_dataloader
from evaluator import MNISTEvaluator
from logger import PythonLogger
from trainer import Trainer
from models import SimpleConvNet, ReBiasModels
class MNISTTrainer(Trainer):
def _set_models(self):
if not self.options.f_config:
self.options.f_config = {'kernel_size': 7, 'feature_pos': 'post'}
self.options.g_config = {'kernel_size': 1, 'feature_pos': 'post'}
f_net = SimpleConvNet(**self.options.f_config)
g_nets = [SimpleConvNet(**self.options.g_config) for _ in range(self.options.n_g_nets)]
self.model = ReBiasModels(f_net, g_nets)
self.evaluator = MNISTEvaluator(device=self.device)
def main(root,
batch_size=256,
train_correlation=0.999,
n_confusing_labels=9,
# optimizer config
lr=0.001,
optim='Adam',
n_epochs=80,
lr_step_size=20,
n_f_pretrain_epochs=0,
n_g_pretrain_epochs=0,
f_lambda_outer=1,
g_lambda_inner=1,
n_g_update=1,
update_g_cls=True,
# criterion config
outer_criterion='RbfHSIC',
inner_criterion='MinusRbfHSIC',
rbf_sigma_scale_x=1,
rbf_sigma_scale_y=1,
rbf_sigma_x=1,
rbf_sigma_y=1,
update_sigma_per_epoch=False,
hsic_alg='unbiased',
feature_pos='post',
# model configs
n_g_nets=1,
f_kernel_size=7,
g_kernel_size=1,
# others
save_dir='./checkpoints',
):
logger = PythonLogger()
logger.log('preparing train loader...')
tr_loader = get_biased_mnist_dataloader(root, batch_size=batch_size,
data_label_correlation=train_correlation,
n_confusing_labels=n_confusing_labels,
train=True)
logger.log('preparing val loader...')
val_loaders = {}
val_loaders['biased'] = get_biased_mnist_dataloader(root, batch_size=batch_size,
data_label_correlation=1,
n_confusing_labels=n_confusing_labels,
train=False)
val_loaders['rho0'] = get_biased_mnist_dataloader(root, batch_size=batch_size,
data_label_correlation=0,
n_confusing_labels=9,
train=False)
val_loaders['unbiased'] = get_biased_mnist_dataloader(root, batch_size=batch_size,
data_label_correlation=0.1,
n_confusing_labels=9,
train=False)
logger.log('preparing trainer...')
log_step = int(100 * 256 / batch_size)
engine = MNISTTrainer(
outer_criterion=outer_criterion,
inner_criterion=inner_criterion,
outer_criterion_config={'sigma_x': rbf_sigma_x, 'sigma_y': rbf_sigma_y,
'algorithm': hsic_alg},
outer_criterion_detail={'sigma_x_type': rbf_sigma_x,
'sigma_y_type': rbf_sigma_y,
'sigma_x_scale': rbf_sigma_scale_x,
'sigma_y_scale': rbf_sigma_scale_y},
inner_criterion_config={'sigma_x': rbf_sigma_x, 'sigma_y': rbf_sigma_y,
'algorithm': hsic_alg},
inner_criterion_detail={'sigma_x_type': rbf_sigma_x,
'sigma_y_type': rbf_sigma_y,
'sigma_x_scale': rbf_sigma_scale_x,
'sigma_y_scale': rbf_sigma_scale_y},
n_epochs=n_epochs,
n_f_pretrain_epochs=n_f_pretrain_epochs,
n_g_pretrain_epochs=n_g_pretrain_epochs,
f_config={'num_classes': 10, 'kernel_size': f_kernel_size, 'feature_pos': feature_pos},
g_config={'num_classes': 10, 'kernel_size': g_kernel_size, 'feature_pos': feature_pos},
f_lambda_outer=f_lambda_outer,
g_lambda_inner=g_lambda_inner,
n_g_update=n_g_update,
update_g_cls=update_g_cls,
n_g_nets=n_g_nets,
optimizer=optim,
f_optim_config={'lr': lr, 'weight_decay': 1e-4},
g_optim_config={'lr': lr, 'weight_decay': 1e-4},
scheduler='StepLR',
f_scheduler_config={'step_size': lr_step_size},
g_scheduler_config={'step_size': lr_step_size},
train_loader=tr_loader,
log_step=log_step,
logger=logger)
engine.train(tr_loader, val_loaders=val_loaders,
val_epoch_step=1,
update_sigma_per_epoch=update_sigma_per_epoch,
save_dir=save_dir)
if __name__ == '__main__':
fire.Fire(main)