-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathextract_bnn_fashion_data.py
75 lines (53 loc) · 2.25 KB
/
extract_bnn_fashion_data.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
import argparse
import pickle
import numpy as np
def get_args():
parser = argparse.ArgumentParser()
parser.add_argument('--name', type=str, default=None,
choices=['svi', 'sgld', 'sghmc'])
return parser.parse_args()
def main():
args = get_args()
exp_num = 5
save_db = dict()
''' pd: processed; npd: non-processed; sc: scratch '''
for name in ['pd', 'npd', 'sc']:
res_err, del_err, test_err = [], [], []
for kk in ['1k', '2k', '3k', '4k', '5k', '6k']:
res_e, del_e, test_e = [], [], []
for ii in range(1, exp_num+1):
if name == 'sc':
path = './{}/{}/remain/remove-{}-ckpt-log.pkl'.format(args.name, ii, kk)
with open(path, 'rb') as f:
dat = pickle.load(f)
res_e.append(1. - dat['(remain) train_acc'][-1])
del_e.append(1. - dat['forgetted_train_acc'][-1])
test_e.append(1. - dat['test_acc'][-1])
else:
path = './{}/{}/forget/forget-{}-ckpt-log.pkl'.format(args.name, ii, kk)
with open(path, 'rb') as f:
dat = pickle.load(f)
idx = -1
if name == 'npd': idx = 0
res_e.append(1. - dat['remain_train_acc'][idx])
del_e.append(1. - dat['forgetted_train_acc'][idx])
test_e.append(1. - dat['test_acc'][idx])
res_err.append(res_e)
del_err.append(del_e)
test_err.append(test_e)
res_err = np.array(res_err)
del_err = np.array(del_err)
test_err = np.array(test_err)
save_db[name] = dict()
save_db[name]['res_err'] = res_err.mean(axis=1)
save_db[name]['del_err'] = del_err.mean(axis=1)
save_db[name]['test_err'] = test_err.mean(axis=1)
cof = exp_num / (exp_num-1)
save_db[name]['res_err_std'] = res_err.std(axis=1) * cof
save_db[name]['del_err_std'] = del_err.std(axis=1) * cof
save_db[name]['test_err_std'] = test_err.std(axis=1) * cof
with open('save-db-{}.pkl'.format(args.name), 'wb') as f:
pickle.dump(save_db, f)
return
if __name__ == '__main__':
main()