-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
206 lines (165 loc) · 7.48 KB
/
main.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
import numpy as np
import h5py as h5
import json
import torch
from torch import nn
from torch import optim
from torch.optim import lr_scheduler
from torch.utils.data import DataLoader
from opts import parse_opts
from model import generate_model
from dataset import get_data_set
from utils import Logger
from train import train_epoch
from validation import val_epoch
from test import test_epoch
import os.path as osp
import os
from sklearn.model_selection import KFold
os.environ["CUDA_VISIBLE_DEVICES"] = "1"
class EarlyStopping():
def __init__(self):
self.nsteps_similar_loss = 0
self.previous_loss = 9999.0
self.delta_loss = 0.01
def _increment_step(self):
self.nsteps_similar_loss += 1
def _reset(self):
self.nsteps_similar_loss = 0
def eval_loss(self, loss):
if (self.previous_loss - loss) <= self.delta_loss:
self._increment_step()
self.previous_loss = loss
else:
self._reset()
self.previous_loss = loss
def get_nsteps(self):
return self.nsteps_similar_loss
if __name__ == '__main__':
opt = parse_opts()
if opt.resume_path:
opt.resume_path = osp.join(opt.root_path, opt.resume_path)
opt.arch = '{}-{}'.format(opt.model, opt.model_depth)
print(opt)
with open(osp.join(opt.result_path, 'opts.json'), 'w') as opt_file:
json.dump(vars(opt), opt_file)
torch.manual_seed(opt.manual_seed)
measures = ['alff', 'T1', 'degree_centrality_binarize', 'degree_centrality_weighted',
'eigenvector_centrality_binarize', 'eigenvector_centrality_weighted', 'lfcd_binarize', 'lfcd_weighted',
'entropy', 'reho', 'vmhc', 'autocorr', 'falff']
#measures = ['degree_centrality_weighted', 'eigenvector_centrality_weighted', 'lfcd_weighted', 'reho', 'vmhc']
if opt.site_wise_cv and opt.kfold_cv:
print('Warning: Both kfold and site-wise CV specified. Defaulting to site_wise')
opt.kfold_cv = False
# Site-wise classification
hfile = h5.File(osp.join(opt.root_path, opt.data_file))
if opt.abide == 1:
to_select = hfile['summaries'].attrs['ABIDE_I_or_II'] == 1 # choose only ABIDEI
site_id = hfile['summaries'].attrs['SITE_ID'][to_select]
all_sites = np.unique(site_id)
n_subjects = len(hfile['summaries'].attrs['DX_GROUP'][to_select]) # DX_GROUP [0,1] -> [ASD, CON]
# Create indices for doing kfold cross-validation
kf = KFold(n_splits=opt.kfolds, random_state=42, shuffle=True)
train_indices = []
val_indices = []
test_indices = []
fold_names = []
if opt.kfold_cv:
X = np.zeros((n_subjects, 1)) # dummy variable to create indices
for fold_i, (l_train, l_test) in enumerate(kf.split(X)):
val_size = int(len(l_train)*opt.val_frac)
train_indices.append(l_train[:-val_size])
val_indices.append(l_train[-val_size:])
test_indices.append(l_test)
fold_names.append(f'kfold_{fold_i}')
if opt.site_wise_cv:
indices = np.arange(n_subjects)
for s in all_sites:
l_train = indices[site_id != s]
l_test = indices[site_id == s]
val_size = int(len(l_train) * opt.val_frac)
train_indices.append(l_train[:-val_size])
val_indices.append(l_train[-val_size:])
test_indices.append(l_test)
fold_names.append(s.decode()) # .decode converts byte to string, e.g., b'UCLA' -> 'UCLA'
for measure in measures:
for train_idx, val_idx, test_idx, fold_name in zip(train_indices, val_indices, test_indices, fold_names):
model, parameters = generate_model(opt)
print(model)
criterion = nn.CrossEntropyLoss()
if not opt.no_cuda:
criterion = criterion.cuda()
if not opt.no_train:
print('Setting up train_loader')
training_data = get_data_set(opt, train_idx, measure, subset=to_select)
train_loader = DataLoader(
training_data,
batch_size=opt.batch_size,
shuffle=True,
num_workers=opt.n_threads,
pin_memory=True)
train_logger = Logger(
os.path.join(opt.result_path, f'train_{measure}_{fold_name}.log'),
['epoch', 'loss', 'acc', 'lr'])
train_batch_logger = Logger(
os.path.join(opt.result_path, 'train_batch.log'),
['epoch', 'batch', 'iter', 'loss', 'acc', 'lr'])
if opt.nesterov:
dampening = 0
else:
dampening = opt.dampening
optimizer = optim.SGD(
parameters,
lr=opt.learning_rate,
momentum=opt.momentum,
dampening=dampening,
weight_decay=opt.weight_decay,
nesterov=opt.nesterov)
scheduler = lr_scheduler.ReduceLROnPlateau(
optimizer, 'min', patience=opt.lr_patience)
if not opt.no_val:
print('Setting up validation_loader')
validation_data = get_data_set(opt, val_idx, measure, subset=to_select)
val_loader = DataLoader(
validation_data,
batch_size=opt.batch_size,
shuffle=False,
num_workers=opt.n_threads,
pin_memory=True)
val_logger = Logger(
os.path.join(opt.result_path, f'val_{measure}_{fold_name}.log'), ['epoch', 'loss', 'acc'])
if opt.resume_path:
print('loading checkpoint {}'.format(opt.resume_path))
checkpoint = torch.load(opt.resume_path)
assert opt.arch == checkpoint['arch']
opt.begin_epoch = checkpoint['epoch']
model.load_state_dict(checkpoint['state_dict'])
if not opt.no_train:
optimizer.load_state_dict(checkpoint['optimizer'])
print('run')
stop_criterion = EarlyStopping()
for i in range(opt.begin_epoch, opt.n_epochs + 1):
if not opt.no_train:
train_epoch(i, train_loader, model, criterion, optimizer, opt,
train_logger, train_batch_logger)
if not opt.no_val:
validation_loss = val_epoch(i, val_loader, model, criterion, opt,
val_logger)
stop_criterion.eval_loss(validation_loss)
if stop_criterion.get_nsteps() >= 10:
break
if not opt.no_train and not opt.no_val:
scheduler.step(validation_loss)
if not opt.no_test:
print('Setting up test_loader')
test_data = get_data_set(opt, test_idx, measure, subset=to_select)
test_loader = torch.utils.data.DataLoader(
test_data,
batch_size=opt.batch_size,
shuffle=False,
num_workers=opt.n_threads,
pin_memory=True)
test_logger = Logger(
os.path.join(opt.result_path, f'test_{measure}_{fold_name}.log'), ['loss', 'acc'])
test_loss = test_epoch(test_loader, model, criterion, opt,
test_logger)