-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathutils.py
More file actions
207 lines (166 loc) · 5.31 KB
/
utils.py
File metadata and controls
207 lines (166 loc) · 5.31 KB
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
import sys
import os
import torch
import numpy as np
import csv
np.random.seed(10)
class Logger(object):
def __init__(self, fpath=None, mode='w'):
self.console = sys.stdout
self.file = None
if fpath is not None:
self.file = open(fpath, mode)
def __del__(self):
self.close()
def __enter__(self):
pass
def __exit__(self, *args):
self.close()
def write(self, msg):
self.console.write(msg)
if self.file is not None:
self.file.write(msg)
def flush(self):
self.console.flush()
if self.file is not None:
self.file.flush()
os.fsync(self.file.fileno())
def close(self):
self.console.close()
if self.file is not None:
self.file.close()
class CSVBatchLogger:
def __init__(self, csv_path, n_groups, mode='w'):
columns = ['epoch', 'batch']
for idx in range(n_groups):
columns.append(f'avg_loss_group:{idx}')
columns.append(f'exp_avg_loss_group:{idx}')
columns.append(f'avg_acc_group:{idx}')
columns.append(f'processed_data_count_group:{idx}')
columns.append(f'update_data_count_group:{idx}')
columns.append(f'update_batch_count_group:{idx}')
columns.append('avg_actual_loss')
columns.append('avg_per_sample_loss')
columns.append('avg_acc')
columns.append('model_norm_sq')
columns.append('reg_loss')
self.path = csv_path
self.file = open(csv_path, mode)
self.columns = columns
self.writer = csv.DictWriter(self.file, fieldnames=columns)
if mode=='w':
self.writer.writeheader()
def log(self, epoch, batch, stats_dict):
stats_dict['epoch'] = epoch
stats_dict['batch'] = batch
self.writer.writerow(stats_dict)
def flush(self):
self.file.flush()
def close(self):
self.file.close()
class AverageMeter(object):
"""Computes and stores the average and current value"""
def __init__(self):
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
def accuracy(output, target, topk=(1,)):
"""Computes the precision@k for the specified values of k"""
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
temp = target.view(1, -1).expand_as(pred)
temp = temp.cuda()
correct = pred.eq(temp)
res = []
for k in topk:
correct_k = correct[:k].view(-1).float().sum(0, keepdim=True)
res.append(correct_k.mul_(100.0 / batch_size))
return res
def set_seed(seed):
"""Sets seed"""
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
torch.manual_seed(seed)
np.random.seed(seed)
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = True
def log_args(args, logger):
for argname, argval in vars(args).items():
logger.write(f'{argname.replace("_"," ").capitalize()}: {argval}\n')
logger.write('\n')
def get_balanced_data(train_data, dim=1):
n = len(train_data)
print('n=',n)
#print(groups)
#for i in range(100):
# print(train_data[i][1],train_data[i][2])
idx = []
for i in range(n):
idx.append(train_data[i][dim])
idx = np.asarray(idx)
print(idx)
n0=np.where(idx==0)[0]
n1=np.where(idx==1)[0]
if dim==2:
n3=np.where(idx==3)[0]
n2=np.where(idx==2)[0]
#n0=min(n0,np.shape(n3)[0])
#n1=min(n1,np.shape(n2)[0])
print('n2 = ',len(n2),', n3 = ',len(n3))
#print(max(n2),max(n3))
print('n0 = ',len(n0),', n1 = ',len(n1))
#print(max(n0),max(n1))
n=min(np.shape(n0)[0],np.shape(n1)[0],20000)
if dim==2:
n=min(n,np.shape(n3)[0],np.shape(n2)[0])
a0=np.random.choice(n0,n,replace=False)
a1=np.random.choice(n1,n,replace=False)
a=np.concatenate((a0,a1))
if dim==2:
a3=np.random.choice(n3,n,replace=False)
a2=np.random.choice(n2,n,replace=False)
a=np.concatenate((a0,a1,a2,a3))
new_data = []
for i in a:
new_data.append(train_data[i])
return new_data
def get_balanced_data_gen(train_data, dim=1, args=None):
n = len(train_data)
print('n=',n)
#print(groups)
#for i in range(100):
# print(train_data[i][1],train_data[i][2])
idx = []
for i in range(n):
idx.append(train_data[i][dim])
idx = np.asarray(idx)
print(idx)
dict1={}
num=[]
for i in np.unique(idx):
nn=np.where(idx==i)[0]
dict1['n'+str(i)]=nn
num.append(len(nn))
nn=min(num)
for i in np.unique(idx):
if i==0:
idx2=np.random.choice(dict1['n'+str(i)],nn,replace=False)
else:
idx2=np.concatenate((idx2,np.random.choice(dict1['n'+str(i)],nn,replace=False)))
new_data = []
for i in idx2:
if args.model.startswith('bert'):
new_data.append((train_data[i][0][:,0], train_data[i][1], train_data[i][2]))
else:
new_data.append(train_data[i])
return new_data